### What changes were proposed in this pull request?
Refactor ScriptTransformation to remove input parameter and replace it by child.output
### Why are the changes needed?
refactor code
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existed UT
Closes#32228 from AngersZhuuuu/SPARK-34035.
Lead-authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Co-authored-by: AngersZhuuuu <angers.zhu@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR implements the decorrelation technique in the paper "Unnesting Arbitrary Queries" by T. Neumann; A. Kemper
(http://www.btw-2015.de/res/proceedings/Hauptband/Wiss/Neumann-Unnesting_Arbitrary_Querie.pdf). It currently supports Filter, Project, Aggregate, Join, and UnaryNode that passes CheckAnalysis.
This feature can be controlled by the config `spark.sql.optimizer.decorrelateInnerQuery.enabled` (default: true).
A few notes:
1. This PR does not relax any constraints in CheckAnalysis for correlated subqueries, even though some cases can be supported by this new framework, such as aggregate with correlated non-equality predicates. This PR focuses on adding the new framework and making sure all existing cases can be supported. Constraints can be relaxed gradually in the future via separate PRs.
2. The new framework is only enabled for correlated scalar subqueries, as the first step. EXISTS/IN subqueries can be supported in the future.
### Why are the changes needed?
Currently, Spark has limited support for correlated subqueries. It only allows `Filter` to reference outer query columns and does not support non-equality predicates when the subquery is aggregated. This new framework will allow more operators to host outer column references and support correlated non-equality predicates and more types of operators in correlated subqueries.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existing unit and SQL query tests and new optimizer plan tests.
Closes#32072 from allisonwang-db/spark-34974-decorrelation.
Authored-by: allisonwang-db <66282705+allisonwang-db@users.noreply.github.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
- Share a static ImmutableBitSet for `treePatternBits` in all object instances of AttributeReference.
- Share three static ImmutableBitSets for `treePatternBits` in three kinds of Literals.
- Add an ImmutableBitSet as a subclass of BitSet.
### Why are the changes needed?
Reduce the additional memory usage caused by `treePatternBits`.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Existing tests.
Closes#32157 from sigmod/leaf.
Authored-by: Yingyi Bu <yingyi.bu@databricks.com>
Signed-off-by: Gengliang Wang <ltnwgl@gmail.com>
### What changes were proposed in this pull request?
This PR updated the `foundNonEqualCorrelatedPred` logic for correlated subqueries in `CheckAnalysis` to only allow correlated equality predicates that guarantee one-to-one mapping between inner and outer attributes, instead of all equality predicates.
### Why are the changes needed?
To fix correctness bugs. Before this fix Spark can give wrong results for certain correlated subqueries that pass CheckAnalysis:
Example 1:
```sql
create or replace view t1(c) as values ('a'), ('b')
create or replace view t2(c) as values ('ab'), ('abc'), ('bc')
select c, (select count(*) from t2 where t1.c = substring(t2.c, 1, 1)) from t1
```
Correct results: [(a, 2), (b, 1)]
Spark results:
```
+---+-----------------+
|c |scalarsubquery(c)|
+---+-----------------+
|a |1 |
|a |1 |
|b |1 |
+---+-----------------+
```
Example 2:
```sql
create or replace view t1(a, b) as values (0, 6), (1, 5), (2, 4), (3, 3);
create or replace view t2(c) as values (6);
select c, (select count(*) from t1 where a + b = c) from t2;
```
Correct results: [(6, 4)]
Spark results:
```
+---+-----------------+
|c |scalarsubquery(c)|
+---+-----------------+
|6 |1 |
|6 |1 |
|6 |1 |
|6 |1 |
+---+-----------------+
```
### Does this PR introduce _any_ user-facing change?
Yes. Users will not be able to run queries that contain unsupported correlated equality predicates.
### How was this patch tested?
Added unit tests.
Closes#32179 from allisonwang-db/spark-35080-subquery-bug.
Lead-authored-by: allisonwang-db <66282705+allisonwang-db@users.noreply.github.com>
Co-authored-by: Wenchen Fan <cloud0fan@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
## What changes were proposed in this pull request?
This PR fixes a couple of things in TypeCoercion rules:
- Only run the propagate types step if the children of a node have output attributes with changed dataTypes and/or nullability. This is implemented as custom tree transformation. The TypeCoercion rules now only implement a partial function.
- Combine multiple type coercion rules into a single rule. Multiple rules are applied in single tree traversal.
- Reduce calls to conf.get in DecimalPrecision. This now happens once per tree traversal, instead of once per matched expression.
- Reduce the use of withNewChildren.
This brings down the number of CPU cycles spend in analysis by ~28% (benchmark: 10 iterations of all TPC-DS queries on SF10).
## How was this patch tested?
Existing tests.
Closes#32208 from sigmod/coercion.
Authored-by: Yingyi Bu <yingyi.bu@databricks.com>
Signed-off-by: herman <herman@databricks.com>
### What changes were proposed in this pull request?
Remove Antlr 4.7 workaround.
### Why are the changes needed?
The https://github.com/antlr/antlr4/commit/ac9f7530 has been fixed in upstream, so remove the workaround to simplify code.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existed UTs.
Closes#32238 from pan3793/antlr-minor.
Authored-by: Cheng Pan <379377944@qq.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Similarly to the test from the PR https://github.com/apache/spark/pull/31799, add tests:
1. Months -> Period -> Months
2. Period -> Months -> Period
3. Duration -> micros -> Duration
### Why are the changes needed?
Add round trip tests for period <-> month and duration <-> micros
### Does this PR introduce _any_ user-facing change?
'No'. Just test cases.
### How was this patch tested?
Jenkins test
Closes#32234 from beliefer/SPARK-34715.
Authored-by: gengjiaan <gengjiaan@360.cn>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Parse the year-month interval literals like `INTERVAL '1-1' YEAR TO MONTH` to values of `YearMonthIntervalType`, and day-time interval literals to `DayTimeIntervalType` values. Currently, Spark SQL supports:
- DAY TO HOUR
- DAY TO MINUTE
- DAY TO SECOND
- HOUR TO MINUTE
- HOUR TO SECOND
- MINUTE TO SECOND
All such interval literals are converted to `DayTimeIntervalType`, and `YEAR TO MONTH` to `YearMonthIntervalType` while loosing info about `from` and `to` units.
**Note**: new behavior is under the SQL config `spark.sql.legacy.interval.enabled` which is `false` by default. When the config is set to `true`, the interval literals are parsed to `CaledarIntervalType` values.
Closes#32176
### Why are the changes needed?
To conform the ANSI SQL standard which assumes conversions of interval literals to year-month or day-time interval but not to mixed interval type like Catalyst's `CalendarIntervalType`.
### Does this PR introduce _any_ user-facing change?
Yes.
Before:
```sql
spark-sql> SELECT INTERVAL '1 01:02:03.123' DAY TO SECOND;
1 days 1 hours 2 minutes 3.123 seconds
spark-sql> SELECT typeof(INTERVAL '1 01:02:03.123' DAY TO SECOND);
interval
```
After:
```sql
spark-sql> SELECT INTERVAL '1 01:02:03.123' DAY TO SECOND;
1 01:02:03.123000000
spark-sql> SELECT typeof(INTERVAL '1 01:02:03.123' DAY TO SECOND);
day-time interval
```
### How was this patch tested?
1. By running the affected test suites:
```
$ ./build/sbt "test:testOnly *.ExpressionParserSuite"
$ SPARK_GENERATE_GOLDEN_FILES=1 build/sbt "sql/testOnly *SQLQueryTestSuite -- -z interval.sql"
$ SPARK_GENERATE_GOLDEN_FILES=1 build/sbt "sql/testOnly *SQLQueryTestSuite -- -z create_view.sql"
$ SPARK_GENERATE_GOLDEN_FILES=1 build/sbt "sql/testOnly *SQLQueryTestSuite -- -z date.sql"
$ SPARK_GENERATE_GOLDEN_FILES=1 build/sbt "sql/testOnly *SQLQueryTestSuite -- -z timestamp.sql"
```
2. PostgresSQL tests are executed with `spark.sql.legacy.interval.enabled` is set to `true` to keep compatibility with PostgreSQL output:
```sql
> SELECT interval '999' second;
0 years 0 mons 0 days 0 hours 16 mins 39.00 secs
```
Closes#32209 from MaxGekk/parse-ansi-interval-literals.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Extend the `Average` expression to support `DayTimeIntervalType` and `YearMonthIntervalType` added by #31614.
Note: the expressions can throw the overflow exception independently from the SQL config `spark.sql.ansi.enabled`. In this way, the modified expressions always behave in the ANSI mode for the intervals.
### Why are the changes needed?
Extend `org.apache.spark.sql.catalyst.expressions.aggregate.Average` to support `DayTimeIntervalType` and `YearMonthIntervalType`.
### Does this PR introduce _any_ user-facing change?
'No'.
Should not since new types have not been released yet.
### How was this patch tested?
Jenkins test
Closes#32229 from beliefer/SPARK-34837.
Authored-by: gengjiaan <gengjiaan@360.cn>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
This PR makes the input buffer configurable (as an internal configuration). This is mainly to work around the regression in uniVocity/univocity-parsers#449.
This is particularly useful for SQL workloads that requires to rewrite the `CREATE TABLE` with options.
### Why are the changes needed?
To work around uniVocity/univocity-parsers#449.
### Does this PR introduce _any_ user-facing change?
No, it's only internal option.
### How was this patch tested?
Manually tested by modifying the unittest added in https://github.com/apache/spark/pull/31858 as below:
```diff
diff --git a/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVSuite.scala
index fd25a79619d..705f38dbfbd 100644
--- a/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVSuite.scala
+++ b/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVSuite.scala
-2456,6 +2456,7 abstract class CSVSuite
test("SPARK-34768: counting a long record with ignoreTrailingWhiteSpace set to true") {
val bufSize = 128
val line = "X" * (bufSize - 1) + "| |"
+ spark.conf.set("spark.sql.csv.parser.inputBufferSize", 128)
withTempPath { path =>
Seq(line).toDF.write.text(path.getAbsolutePath)
assert(spark.read.format("csv")
```
Closes#32231 from HyukjinKwon/SPARK-35045-followup.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Now that `AnalysisOnlyCommand` in introduced in #32032, `CacheTable` and `UncacheTable` can extend `AnalysisOnlyCommand` to simplify the code base. For example, the logic to handle these commands such that the tables are only analyzed is scattered across different places.
### Why are the changes needed?
To simplify the code base to handle these two commands.
### Does this PR introduce _any_ user-facing change?
No, just internal refactoring.
### How was this patch tested?
The existing tests (e.g., `CachedTableSuite`) cover the changes in this PR. For example, if I make `CacheTable`/`UncacheTable` extend `LeafCommand`, there are few failures in `CachedTableSuite`.
Closes#32220 from imback82/cache_cmd_analysis_only.
Authored-by: Terry Kim <yuminkim@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR:
- Adds a new expression `GroupingExprRef` that can be used in aggregate expressions of `Aggregate` nodes to refer grouping expressions by index. These expressions capture the data type and nullability of the referred grouping expression.
- Adds a new rule `EnforceGroupingReferencesInAggregates` that inserts the references in the beginning of the optimization phase.
- Adds a new rule `UpdateGroupingExprRefNullability` to update nullability of `GroupingExprRef` expressions as nullability of referred grouping expression can change during optimization.
### Why are the changes needed?
If aggregate expressions (without aggregate functions) in an `Aggregate` node are complex then the `Optimizer` can optimize out grouping expressions from them and so making aggregate expressions invalid.
Here is a simple example:
```
SELECT not(t.id IS NULL) , count(*)
FROM t
GROUP BY t.id IS NULL
```
In this case the `BooleanSimplification` rule does this:
```
=== Applying Rule org.apache.spark.sql.catalyst.optimizer.BooleanSimplification ===
!Aggregate [isnull(id#222)], [NOT isnull(id#222) AS (NOT (id IS NULL))#226, count(1) AS c#224L] Aggregate [isnull(id#222)], [isnotnull(id#222) AS (NOT (id IS NULL))#226, count(1) AS c#224L]
+- Project [value#219 AS id#222] +- Project [value#219 AS id#222]
+- LocalRelation [value#219] +- LocalRelation [value#219]
```
where `NOT isnull(id#222)` is optimized to `isnotnull(id#222)` and so it no longer refers to any grouping expression.
Before this PR:
```
== Optimized Logical Plan ==
Aggregate [isnull(id#222)], [isnotnull(id#222) AS (NOT (id IS NULL))#234, count(1) AS c#232L]
+- Project [value#219 AS id#222]
+- LocalRelation [value#219]
```
and running the query throws an error:
```
Couldn't find id#222 in [isnull(id#222)#230,count(1)#226L]
java.lang.IllegalStateException: Couldn't find id#222 in [isnull(id#222)#230,count(1)#226L]
```
After this PR:
```
== Optimized Logical Plan ==
Aggregate [isnull(id#222)], [NOT groupingexprref(0) AS (NOT (id IS NULL))#234, count(1) AS c#232L]
+- Project [value#219 AS id#222]
+- LocalRelation [value#219]
```
and the query works.
### Does this PR introduce _any_ user-facing change?
Yes, the query works.
### How was this patch tested?
Added new UT.
Closes#31913 from peter-toth/SPARK-34581-keep-grouping-expressions.
Authored-by: Peter Toth <peter.toth@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
It seems that we miss classifying one `SparkOutOfMemoryError` in `HashedRelation`. Add the error classification for it. In addition, clean up two errors definition of `HashJoin` as they are not used.
### Why are the changes needed?
Better error classification.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Existing tests.
Closes#32211 from c21/error-message.
Authored-by: Cheng Su <chengsu@fb.com>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
Add checks for `YearMonthIntervalType` and `DayTimeIntervalType` to `MutableProjectionSuite`.
### Why are the changes needed?
To improve test coverage, and the same checks as for `CalendarIntervalType`.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
By running the modified test suite:
```
$ build/sbt "test:testOnly *MutableProjectionSuite"
```
Closes#32225 from MaxGekk/test-ansi-intervals-in-MutableProjectionSuite.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
Extend the `Sum` expression to to support `DayTimeIntervalType` and `YearMonthIntervalType` added by #31614.
Note: the expressions can throw the overflow exception independently from the SQL config `spark.sql.ansi.enabled`. In this way, the modified expressions always behave in the ANSI mode for the intervals.
### Why are the changes needed?
Extend `org.apache.spark.sql.catalyst.expressions.aggregate.Sum` to support `DayTimeIntervalType` and `YearMonthIntervalType`.
### Does this PR introduce _any_ user-facing change?
'No'.
Should not since new types have not been released yet.
### How was this patch tested?
Jenkins test
Closes#32107 from beliefer/SPARK-34716.
Lead-authored-by: gengjiaan <gengjiaan@360.cn>
Co-authored-by: beliefer <beliefer@163.com>
Co-authored-by: Hyukjin Kwon <gurwls223@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
In the PR, I propose to add additional checks for ANSI interval types `YearMonthIntervalType` and `DayTimeIntervalType` to `LiteralExpressionSuite`.
Also, I replaced some long literal values by `CalendarInterval` to check `CalendarIntervalType` that the tests were supposed to check.
### Why are the changes needed?
To improve test coverage and have the same checks for ANSI types as for `CalendarIntervalType`.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
By running the modified test suite:
```
$ build/sbt "test:testOnly *LiteralExpressionSuite"
```
Closes#32213 from MaxGekk/interval-literal-tests.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
The precision of `java.time.Duration` is nanosecond, but when it is used as `DayTimeIntervalType` in Spark, it is microsecond.
At present, the `DayTimeIntervalType` data generated in the implementation of `RandomDataGenerator` is accurate to nanosecond, which will cause the `DayTimeIntervalType` to be converted to long, and then back to `DayTimeIntervalType` to lose the accuracy, which will cause the test to fail. For example: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/137390/testReport/org.apache.spark.sql.hive.execution/HashAggregationQueryWithControlledFallbackSuite/udaf_with_all_data_types/
### Why are the changes needed?
Improve `RandomDataGenerator` so that the generated data fits the precision of DayTimeIntervalType in spark.
### Does this PR introduce _any_ user-facing change?
'No'. Just change the test class.
### How was this patch tested?
Jenkins test.
Closes#32212 from beliefer/SPARK-35116.
Authored-by: beliefer <beliefer@163.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
This issue fixes an issue that indentation of multiple output JSON records in a single split file are broken except for the first record in the split when `pretty` option is `true`.
```
// Run in the Spark Shell.
// Set spark.sql.leafNodeDefaultParallelism to 1 for the current master.
// Or set spark.default.parallelism for the previous releases.
spark.conf.set("spark.sql.leafNodeDefaultParallelism", 1)
val df = Seq("a", "b", "c").toDF
df.write.option("pretty", "true").json("/path/to/output")
# Run in a Shell
$ cat /path/to/output/*.json
{
"value" : "a"
}
{
"value" : "b"
}
{
"value" : "c"
}
```
### Why are the changes needed?
It's not pretty even though `pretty` option is true.
### Does this PR introduce _any_ user-facing change?
I think "No". Indentation style is changed but JSON format is not changed.
### How was this patch tested?
New test.
Closes#32203 from sarutak/fix-ugly-indentation.
Authored-by: Kousuke Saruta <sarutak@oss.nttdata.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Handle `YearMonthIntervalType` and `DayTimeIntervalType` in the `sql()` and `toString()` method of `Literal`, and format the ANSI interval in the ANSI style.
### Why are the changes needed?
To improve readability and UX with Spark SQL. For example, a test output before the changes:
```
-- !query
select timestamp'2011-11-11 11:11:11' - interval '2' day
-- !query schema
struct<TIMESTAMP '2011-11-11 11:11:11' - 172800000000:timestamp>
-- !query output
2011-11-09 11:11:11
```
### Does this PR introduce _any_ user-facing change?
Should not since the new intervals haven't been released yet.
### How was this patch tested?
By running new tests:
```
$ ./build/sbt "test:testOnly *LiteralExpressionSuite"
```
Closes#32196 from MaxGekk/literal-ansi-interval-sql.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Normal function parameters should not support alias, hive not support too
In this pr we forbid use alias in `TRANSFORM`'s inputs
### Why are the changes needed?
Fix bug
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Added UT
Closes#32165 from AngersZhuuuu/SPARK-35070.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Support `date +/- day-time interval`. In the PR, I propose to update the binary arithmetic rules, and cast an input date to a timestamp at the session time zone, and then add a day-time interval to it.
### Why are the changes needed?
1. To conform the ANSI SQL standard which requires to support such operation over dates and intervals:
<img width="811" alt="Screenshot 2021-03-12 at 11 36 14" src="https://user-images.githubusercontent.com/1580697/111081674-865d4900-8515-11eb-86c8-3538ecaf4804.png">
2. To fix the regression comparing to the recent Spark release 3.1 with default settings.
Before the changes:
```sql
spark-sql> select date'now' + (timestamp'now' - timestamp'yesterday');
Error in query: cannot resolve 'DATE '2021-04-14' + subtracttimestamps(TIMESTAMP '2021-04-14 18:14:56.497', TIMESTAMP '2021-04-13 00:00:00')' due to data type mismatch: argument 1 requires timestamp type, however, 'DATE '2021-04-14'' is of date type.; line 1 pos 7;
'Project [unresolvedalias(cast(2021-04-14 + subtracttimestamps(2021-04-14 18:14:56.497, 2021-04-13 00:00:00, false, Some(Europe/Moscow)) as date), None)]
+- OneRowRelation
```
Spark 3.1:
```sql
spark-sql> select date'now' + (timestamp'now' - timestamp'yesterday');
2021-04-15
```
Hive:
```sql
0: jdbc:hive2://localhost:10000/default> select date'2021-04-14' + (timestamp'2020-04-14 18:15:30' - timestamp'2020-04-13 00:00:00');
+------------------------+
| _c0 |
+------------------------+
| 2021-04-15 18:15:30.0 |
+------------------------+
```
### Does this PR introduce _any_ user-facing change?
Should not since new intervals have not been released yet.
After the changes:
```sql
spark-sql> select date'now' + (timestamp'now' - timestamp'yesterday');
2021-04-15 18:13:16.555
```
### How was this patch tested?
By running new tests:
```
$ build/sbt "test:testOnly *ColumnExpressionSuite"
```
Closes#32170 from MaxGekk/date-add-day-time-interval.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
According to https://github.com/apache/spark/pull/29087#discussion_r612267050, add UT in `transform.sql`
It seems that distinct is not recognized as a reserved word here
```
-- !query
explain extended SELECT TRANSFORM(distinct b, a, c)
USING 'cat' AS (a, b, c)
FROM script_trans
WHERE a <= 4
-- !query schema
struct<plan:string>
-- !query output
== Parsed Logical Plan ==
'ScriptTransformation [*], cat, [a#x, b#x, c#x], ScriptInputOutputSchema(List(),List(),None,None,List(),List(),None,None,false)
+- 'Project ['distinct AS b#x, 'a, 'c]
+- 'Filter ('a <= 4)
+- 'UnresolvedRelation [script_trans], [], false
== Analyzed Logical Plan ==
org.apache.spark.sql.AnalysisException: cannot resolve 'distinct' given input columns: [script_trans.a, script_trans.b, script_trans.c]; line 1 pos 34;
'ScriptTransformation [*], cat, [a#x, b#x, c#x], ScriptInputOutputSchema(List(),List(),None,None,List(),List(),None,None,false)
+- 'Project ['distinct AS b#x, a#x, c#x]
+- Filter (a#x <= 4)
+- SubqueryAlias script_trans
+- View (`script_trans`, [a#x,b#x,c#x])
+- Project [cast(a#x as int) AS a#x, cast(b#x as int) AS b#x, cast(c#x as int) AS c#x]
+- Project [a#x, b#x, c#x]
+- SubqueryAlias script_trans
+- LocalRelation [a#x, b#x, c#x]
```
Hive's error
### Why are the changes needed?
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Added Ut
Closes#32149 from AngersZhuuuu/SPARK-28227-new-followup.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR proposes to introduce the `AnalysisOnlyCommand` trait such that a command that extends this trait can have its children only analyzed, but not optimized. There is a corresponding analysis rule `HandleAnalysisOnlyCommand` that marks the command as analyzed after all other analysis rules are run.
This can be useful if a logical plan has children where they need to be only analyzed, but not optimized - e.g., `CREATE VIEW` or `CACHE TABLE AS`. This also addresses the issue found in #31933.
This PR also updates `CreateViewCommand`, `CacheTableAsSelect`, and `AlterViewAsCommand` to use the new trait / rule such that their children are only analyzed.
### Why are the changes needed?
To address the issue where the plan is unnecessarily re-analyzed in `CreateViewCommand`.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Existing tests should cover the changes.
Closes#32032 from imback82/skip_transform.
Authored-by: Terry Kim <yuminkim@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Adds the duplicated common columns as hidden columns to the Projection used to rewrite NATURAL/USING JOINs.
### Why are the changes needed?
Allows users to resolve either side of the NATURAL/USING JOIN's common keys.
Previously, the user could only resolve the following columns:
| Join type | Left key columns | Right key columns |
| --- | --- | --- |
| Inner | Yes | No |
| Left | Yes | No |
| Right | No | Yes |
| Outer | No | No |
### Does this PR introduce _any_ user-facing change?
Yes. The user can now symmetrically resolve the common columns from a NATURAL/USING JOIN.
### How was this patch tested?
SQL-side tests. The behavior matches PostgreSQL and MySQL.
Closes#31666 from karenfeng/spark-34527.
Authored-by: Karen Feng <karen.feng@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR fixes an issue that `LIST FILES/JARS/ARCHIVES path1 path2 ...` cannot list all paths if at least one path is quoted.
An example here.
```
ADD FILE /tmp/test1;
ADD FILE /tmp/test2;
LIST FILES /tmp/test1 /tmp/test2;
file:/tmp/test1
file:/tmp/test2
LIST FILES /tmp/test1 "/tmp/test2";
file:/tmp/test2
```
In this example, the second `LIST FILES` doesn't show `file:/tmp/test1`.
To resolve this issue, I modified the syntax rule to be able to handle this case.
I also changed `SparkSQLParser` to be able to handle paths which contains white spaces.
### Why are the changes needed?
This is a bug.
I also have a plan which extends `ADD FILE/JAR/ARCHIVE` to take multiple paths like Hive and the syntax rule change is necessary for that.
### Does this PR introduce _any_ user-facing change?
Yes. Users can pass quoted paths when using `ADD FILE/JAR/ARCHIVE`.
### How was this patch tested?
New test.
Closes#32074 from sarutak/fix-list-files-bug.
Authored-by: Kousuke Saruta <sarutak@oss.nttdata.com>
Signed-off-by: Kousuke Saruta <sarutak@oss.nttdata.com>
### What changes were proposed in this pull request?
This PR group exception messages in `/core/src/main/scala/org/apache/spark/sql/execution`.
### Why are the changes needed?
It will largely help with standardization of error messages and its maintenance.
### Does this PR introduce _any_ user-facing change?
No. Error messages remain unchanged.
### How was this patch tested?
No new tests - pass all original tests to make sure it doesn't break any existing behavior.
Closes#31920 from beliefer/SPARK-33604.
Lead-authored-by: gengjiaan <gengjiaan@360.cn>
Co-authored-by: Jiaan Geng <beliefer@163.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Populate table catalog and identifier from `DataStreamWriter` to `WriteToMicroBatchDataSource` so that we can invalidate cache for tables that are updated by a streaming write.
This is somewhat related [SPARK-27484](https://issues.apache.org/jira/browse/SPARK-27484) and [SPARK-34183](https://issues.apache.org/jira/browse/SPARK-34183) (#31700), as ideally we may want to replace `WriteToMicroBatchDataSource` and `WriteToDataSourceV2` with logical write nodes and feed them to analyzer. That will potentially change the code path involved in this PR.
### Why are the changes needed?
Currently `WriteToDataSourceV2` doesn't have cache invalidation logic, and therefore, when the target table for a micro batch streaming job is cached, the cache entry won't be removed when the table is updated.
### Does this PR introduce _any_ user-facing change?
Yes now when a DSv2 table which supports streaming write is updated by a streaming job, its cache will also be invalidated.
### How was this patch tested?
Added a new UT.
Closes#32039 from sunchao/streaming-cache.
Authored-by: Chao Sun <sunchao@apple.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Remove duplicate code in `TreeNode.treePatternBits`
### Why are the changes needed?
Code clean up. Make it easier for maintainence.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existing tests.
Closes#32143 from gengliangwang/getBits.
Authored-by: Gengliang Wang <ltnwgl@gmail.com>
Signed-off-by: Gengliang Wang <ltnwgl@gmail.com>
### What changes were proposed in this pull request?
This PR makes the input buffer configurable (as an internal option). This is mainly to work around uniVocity/univocity-parsers#449.
### Why are the changes needed?
To work around uniVocity/univocity-parsers#449.
### Does this PR introduce _any_ user-facing change?
No, it's only internal option.
### How was this patch tested?
Manually tested by modifying the unittest added in https://github.com/apache/spark/pull/31858 as below:
```diff
diff --git a/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVSuite.scala
index fd25a79619d..b58f0bd3661 100644
--- a/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVSuite.scala
+++ b/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVSuite.scala
-2460,6 +2460,7 abstract class CSVSuite
Seq(line).toDF.write.text(path.getAbsolutePath)
assert(spark.read.format("csv")
.option("delimiter", "|")
+ .option("inputBufferSize", "128")
.option("ignoreTrailingWhiteSpace", "true").load(path.getAbsolutePath).count() == 1)
}
}
```
Closes#32145 from HyukjinKwon/SPARK-35045.
Lead-authored-by: Hyukjin Kwon <gurwls223@apache.org>
Co-authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Fix PhysicalAggregation to not transform a foldable expression.
### Why are the changes needed?
It can potentially break certain queries like the added unit test shows.
### Does this PR introduce _any_ user-facing change?
Yes, it fixes undesirable errors caused by a returned TypeCheckFailure from places like RegExpReplace.checkInputDataTypes.
Closes#32113 from sigmod/foldable.
Authored-by: Yingyi Bu <yingyi.bu@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR contains:
- AnalysisHelper changes to allow the resolve function family to stop earlier without traversing the entire tree;
- Example changes in a few rules to support such pruning, e.g., ResolveRandomSeed, ResolveWindowFrame, ResolveWindowOrder, and ResolveNaturalAndUsingJoin.
### Why are the changes needed?
It's a framework-level change for reducing the query compilation time.
In particular, if we update existing analysis rules' call sites as per the examples in this PR, the analysis time can be reduced as described in the [doc](https://docs.google.com/document/d/1SEUhkbo8X-0cYAJFYFDQhxUnKJBz4lLn3u4xR2qfWqk).
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
It is tested by existing tests.
Closes#32135 from sigmod/resolver.
Authored-by: Yingyi Bu <yingyi.bu@databricks.com>
Signed-off-by: Gengliang Wang <ltnwgl@gmail.com>
### What changes were proposed in this pull request?
This PR allows non-aggregated correlated scalar subquery if the max output row is less than 2. Correlated scalar subqueries need to be aggregated because they are going to be decorrelated and rewritten as LEFT OUTER joins. If the correlated scalar subquery produces more than one output row, the rewrite will yield wrong results.
But this constraint can be relaxed when the subquery plan's the max number of output rows is less than or equal to 1.
### Why are the changes needed?
To relax a constraint in CheckAnalysis for the correlated scalar subquery.
### Does this PR introduce _any_ user-facing change?
Yes
### How was this patch tested?
Unit tests
Closes#32111 from allisonwang-db/spark-28379-aggregated.
Authored-by: allisonwang-db <66282705+allisonwang-db@users.noreply.github.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Supports cardinality estimation of union, sort and range operator.
1. **Union**: number of rows in output will be the sum of number of rows in the output for each child of union, min and max for each column in the output will be the min and max of that particular column coming from its children.
Example:
Table 1
a b
1 6
2 3
Table 2
a b
1 3
4 1
stats for table1 union table2 would be number of rows = 4, columnStats = (a: {min: 1, max: 4}, b: {min: 1, max: 6})
2. **Sort**: row and columns stats would be same as its children.
3. **Range**: number of output rows and distinct count will be equal to number of elements, min and max is calculated from start, end and step param.
### Why are the changes needed?
The change will enhance the feature https://issues.apache.org/jira/browse/SPARK-16026 and will help in other stats based optimizations.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
New unit tests added.
Closes#30334 from ayushi-agarwal/SPARK-33411.
Lead-authored-by: ayushi agarwal <ayaga@microsoft.com>
Co-authored-by: ayushi-agarwal <36420535+ayushi-agarwal@users.noreply.github.com>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
1. Extend SQL syntax rules to support a sign before the interval strings of ANSI year-month and day-time intervals.
2. Recognize `-` in `AstBuilder` and negate parsed intervals.
### Why are the changes needed?
To conform to the SQL standard which allows a sign before the string interval, see `"5.3 <literal>"`:
```
<interval literal> ::=
INTERVAL [ <sign> ] <interval string> <interval qualifier>
<interval string> ::=
<quote> <unquoted interval string> <quote>
<unquoted interval string> ::=
[ <sign> ] { <year-month literal> | <day-time literal> }
<sign> ::=
<plus sign>
| <minus sign>
```
### Does this PR introduce _any_ user-facing change?
Should not because it just extends supported intervals syntax.
### How was this patch tested?
By running new tests in `interval.sql`:
```
$ build/sbt "sql/testOnly *SQLQueryTestSuite -- -z interval.sql"
```
Closes#32134 from MaxGekk/negative-parsed-intervals.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Disallow group by aliases under ANSI mode.
### Why are the changes needed?
As per the ANSI SQL standard secion 7.12 <group by clause>:
>Each `grouping column reference` shall unambiguously reference a column of the table resulting from the `from clause`. A column referenced in a `group by clause` is a grouping column.
By forbidding it, we can avoid ambiguous SQL queries like:
```
SELECT col + 1 as col FROM t GROUP BY col
```
### Does this PR introduce _any_ user-facing change?
Yes, group by aliases is not allowed under ANSI mode.
### How was this patch tested?
Unit tests
Closes#32129 from gengliangwang/disallowGroupByAlias.
Authored-by: Gengliang Wang <ltnwgl@gmail.com>
Signed-off-by: Gengliang Wang <ltnwgl@gmail.com>
### What changes were proposed in this pull request?
For Spark SQL, it can't support script transform SQL with aggregationClause/windowClause/LateralView.
This case we can't directly migration Hive SQL to Spark SQL.
In this PR, we treat all script transform statement's query part (exclude transform about part) as a separate query block and solve it as ScriptTransformation's child and pass a UnresolvedStart as ScriptTransform's input. Then in analyzer level, we pass child's output as ScriptTransform's input. Then we can support all kind of normal SELECT query combine with script transformation.
Such as transform with aggregation:
```
SELECT TRANSFORM ( d2, max(d1) as max_d1, sum(d3))
USING 'cat' AS (a,b,c)
FROM script_trans
WHERE d1 <= 100
GROUP BY d2
HAVING max_d1 > 0
```
When we build AST, we treat it as
```
SELECT TRANSFORM (*)
USING 'cat' AS (a,b,c)
FROM (
SELECT d2, max(d1) as max_d1, sum(d3)
FROM script_trans
WHERE d1 <= 100
GROUP BY d2
HAVING max_d1 > 0
) tmp
```
then in Analyzer's `ResolveReferences`, resolve `* (UnresolvedStar)`, then sql behavior like
```
SELECT TRANSFORM ( d2, max(d1) as max_d1, sum(d3))
USING 'cat' AS (a,b,c)
FROM script_trans
WHERE d1 <= 100
GROUP BY d2
HAVING max_d1 > 0
```
About UT, in this pr we add a lot of different SQL to check we can support all kind of such SQL and each kind of expressions can work well, such as alias, case when, binary compute etc...
### Why are the changes needed?
Support transform with aggregateClause/windowClause/LateralView etc , make sql migration more smoothly
### Does this PR introduce _any_ user-facing change?
User can write transform with aggregateClause/windowClause/LateralView.
### How was this patch tested?
Added UT
Closes#29087 from AngersZhuuuu/SPARK-28227-NEW.
Lead-authored-by: angerszhu <angers.zhu@gmail.com>
Co-authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Co-authored-by: AngersZhuuuu <angers.zhu@gmail.com>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
Support GROUP BY use Separate columns and CUBE/ROLLUP
In postgres sql, it support
```
select a, b, c, count(1) from t group by a, b, cube (a, b, c);
select a, b, c, count(1) from t group by a, b, rollup(a, b, c);
select a, b, c, count(1) from t group by cube(a, b), rollup (a, b, c);
select a, b, c, count(1) from t group by a, b, grouping sets((a, b), (a), ());
```
In this pr, we have done two things as below:
1. Support partial grouping analytics such as `group by a, cube(a, b)`
2. Support mixed grouping analytics such as `group by cube(a, b), rollup(b,c)`
*Partial Groupings*
Partial Groupings means there are both `group_expression` and `CUBE|ROLLUP|GROUPING SETS`
in GROUP BY clause. For example:
`GROUP BY warehouse, CUBE(product, location)` is equivalent to
`GROUP BY GROUPING SETS((warehouse, product, location), (warehouse, product), (warehouse, location), (warehouse))`.
`GROUP BY warehouse, ROLLUP(product, location)` is equivalent to
`GROUP BY GROUPING SETS((warehouse, product, location), (warehouse, product), (warehouse))`.
`GROUP BY warehouse, GROUPING SETS((product, location), (producet), ())` is equivalent to
`GROUP BY GROUPING SETS((warehouse, product, location), (warehouse, location), (warehouse))`.
*Concatenated Groupings*
Concatenated groupings offer a concise way to generate useful combinations of groupings. Groupings specified
with concatenated groupings yield the cross-product of groupings from each grouping set. The cross-product
operation enables even a small number of concatenated groupings to generate a large number of final groups.
The concatenated groupings are specified simply by listing multiple `GROUPING SETS`, `CUBES`, and `ROLLUP`,
and separating them with commas. For example:
`GROUP BY GROUPING SETS((warehouse), (producet)), GROUPING SETS((location), (size))` is equivalent to
`GROUP BY GROUPING SETS((warehouse, location), (warehouse, size), (product, location), (product, size))`.
`GROUP BY CUBE((warehouse), (producet)), ROLLUP((location), (size))` is equivalent to
`GROUP BY GROUPING SETS((warehouse, product), (warehouse), (producet), ()), GROUPING SETS((location, size), (location), ())`
`GROUP BY GROUPING SETS(
(warehouse, product, location, size), (warehouse, product, location), (warehouse, product),
(warehouse, location, size), (warehouse, location), (warehouse),
(product, location, size), (product, location), (product),
(location, size), (location), ())`.
`GROUP BY order, CUBE((warehouse), (producet)), ROLLUP((location), (size))` is equivalent to
`GROUP BY order, GROUPING SETS((warehouse, product), (warehouse), (producet), ()), GROUPING SETS((location, size), (location), ())`
`GROUP BY GROUPING SETS(
(order, warehouse, product, location, size), (order, warehouse, product, location), (order, warehouse, product),
(order, warehouse, location, size), (order, warehouse, location), (order, warehouse),
(order, product, location, size), (order, product, location), (order, product),
(order, location, size), (order, location), (order))`.
### Why are the changes needed?
Support more flexible grouping analytics
### Does this PR introduce _any_ user-facing change?
User can use sql like
```
select a, b, c, agg_expr() from table group by a, cube(b, c)
```
### How was this patch tested?
Added UT
Closes#30144 from AngersZhuuuu/SPARK-33229.
Lead-authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Co-authored-by: angerszhu <angers.zhu@gmail.com>
Co-authored-by: Wenchen Fan <cloud0fan@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR contains:
- TreeNode, QueryPlan, AnalysisHelper changes to allow the transform function family to stop earlier without traversing the entire tree;
- Example changes in a few rules to support such pruning, e.g., ReorderJoin and OptimizeIn.
Here is a [design doc](https://docs.google.com/document/d/1SEUhkbo8X-0cYAJFYFDQhxUnKJBz4lLn3u4xR2qfWqk) that elaborates the ideas and benchmark numbers.
### Why are the changes needed?
It's a framework-level change for reducing the query compilation time.
In particular, if we update existing rules and transform call sites as per the examples in this PR, the analysis time and query optimization time can be reduced as described in this [doc](https://docs.google.com/document/d/1SEUhkbo8X-0cYAJFYFDQhxUnKJBz4lLn3u4xR2qfWqk) .
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
It is tested by existing tests.
Closes#32060 from sigmod/bits.
Authored-by: Yingyi Bu <yingyi.bu@databricks.com>
Signed-off-by: Gengliang Wang <ltnwgl@gmail.com>
### What changes were proposed in this pull request?
before when we use aggregate ordinal in group by expression and index position is a aggregate function, it will show error as
```
– !query
select a, b, sum(b) from data group by 3
– !query schema
struct<>
– !query output
org.apache.spark.sql.AnalysisException
aggregate functions are not allowed in GROUP BY, but found sum(data.b)
```
It't not clear enough refactor this error message in this pr
### Why are the changes needed?
refactor error message
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existed UT
Closes#32089 from AngersZhuuuu/SPARK-34986.
Lead-authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Co-authored-by: AngersZhuuuu <angers.zhu@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
1. Extend `IntervalUtils` methods: `toYearMonthIntervalString` and `toDayTimeIntervalString` to support formatting of year-month/day-time intervals in Hive style. The methods get new parameter style which can have to values; `HIVE_STYLE` and `ANSI_STYLE`.
2. Invoke `toYearMonthIntervalString` and `toDayTimeIntervalString` from the `Cast` expression with the `style` parameter is set to `ANSI_STYLE`.
3. Invoke `toYearMonthIntervalString` and `toDayTimeIntervalString` from `HiveResult` with `style` is set to `HIVE_STYLE`.
### Why are the changes needed?
The `spark-sql` shell formats its output in Hive style by using `HiveResult.hiveResultString()`. The changes are needed to match Hive behavior. For instance,
Hive:
```sql
0: jdbc:hive2://localhost:10000/default> select timestamp'2021-01-01 01:02:03.000001' - date'2020-12-31';
+-----------------------+
| _c0 |
+-----------------------+
| 1 01:02:03.000001000 |
+-----------------------+
```
Spark before the changes:
```sql
spark-sql> select timestamp'2021-01-01 01:02:03.000001' - date'2020-12-31';
INTERVAL '1 01:02:03.000001' DAY TO SECOND
```
Also this should unblock #32099 which enables *.sql tests in `SQLQueryTestSuite`.
### Does this PR introduce _any_ user-facing change?
Yes. After the changes:
```sql
spark-sql> select timestamp'2021-01-01 01:02:03.000001' - date'2020-12-31';
1 01:02:03.000001000
```
### How was this patch tested?
1. Added new tests to `IntervalUtilsSuite`:
```
$ build/sbt "test:testOnly *IntervalUtilsSuite"
```
2. Modified existing tests in `HiveResultSuite`:
```
$ build/sbt -Phive-2.3 -Phive-thriftserver "testOnly *HiveResultSuite"
```
3. By running cast tests:
```
$ build/sbt "testOnly *CastSuite*"
```
Closes#32120 from MaxGekk/ansi-intervals-hive-thrift-server.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
This patch proposes a fix of nested column pruning for extracting case-insensitive struct field from array of struct.
### Why are the changes needed?
Under case-insensitive mode, nested column pruning rule cannot correctly push down extractor of a struct field of an array of struct, e.g.,
```scala
val query = spark.table("contacts").select("friends.First", "friends.MiDDle")
```
Error stack:
```
[info] java.lang.IllegalArgumentException: Field "First" does not exist.
[info] Available fields:
[info] at org.apache.spark.sql.types.StructType$$anonfun$apply$1.apply(StructType.scala:274)
[info] at org.apache.spark.sql.types.StructType$$anonfun$apply$1.apply(StructType.scala:274)
[info] at scala.collection.MapLike$class.getOrElse(MapLike.scala:128)
[info] at scala.collection.AbstractMap.getOrElse(Map.scala:59)
[info] at org.apache.spark.sql.types.StructType.apply(StructType.scala:273)
[info] at org.apache.spark.sql.execution.ProjectionOverSchema$$anonfun$getProjection$3.apply(ProjectionOverSchema.scala:44)
[info] at org.apache.spark.sql.execution.ProjectionOverSchema$$anonfun$getProjection$3.apply(ProjectionOverSchema.scala:41)
```
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Unit test
Closes#32059 from viirya/fix-array-nested-pruning.
Authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Signed-off-by: Liang-Chi Hsieh <viirya@gmail.com>
### What changes were proposed in this pull request?
One of the main performance bottlenecks in query compilation is overly-generic tree transformation methods, namely `mapChildren` and `withNewChildren` (defined in `TreeNode`). These methods have an overly-generic implementation to iterate over the children and rely on reflection to create new instances. We have observed that, especially for queries with large query plans, a significant amount of CPU cycles are wasted in these methods. In this PR we make these methods more efficient, by delegating the iteration and instantiation to concrete node types. The benchmarks show that we can expect significant performance improvement in total query compilation time in queries with large query plans (from 30-80%) and about 20% on average.
#### Problem detail
The `mapChildren` method in `TreeNode` is overly generic and costly. To be more specific, this method:
- iterates over all the fields of a node using Scala’s product iterator. While the iteration is not reflection-based, thanks to the Scala compiler generating code for `Product`, we create many anonymous functions and visit many nested structures (recursive calls).
The anonymous functions (presumably compiled to Java anonymous inner classes) also show up quite high on the list in the object allocation profiles, so we are putting unnecessary pressure on GC here.
- does a lot of comparisons. Basically for each element returned from the product iterator, we check if it is a child (contained in the list of children) and then transform it. We can avoid that by just iterating over children, but in the current implementation, we need to gather all the fields (only transform the children) so that we can instantiate the object using the reflection.
- creates objects using reflection, by delegating to the `makeCopy` method, which is several orders of magnitude slower than using the constructor.
#### Solution
The proposed solution in this PR is rather straightforward: we rewrite the `mapChildren` method using the `children` and `withNewChildren` methods. The default `withNewChildren` method suffers from the same problems as `mapChildren` and we need to make it more efficient by specializing it in concrete classes. Similar to how each concrete query plan node already defines its children, it should also define how they can be constructed given a new list of children. Actually, the implementation is quite simple in most cases and is a one-liner thanks to the copy method present in Scala case classes. Note that we cannot abstract over the copy method, it’s generated by the compiler for case classes if no other type higher in the hierarchy defines it. For most concrete nodes, the implementation of `withNewChildren` looks like this:
```
override def withNewChildren(newChildren: Seq[LogicalPlan]): LogicalPlan = copy(children = newChildren)
```
The current `withNewChildren` method has two properties that we should preserve:
- It returns the same instance if the provided children are the same as its children, i.e., it preserves referential equality.
- It copies tags and maintains the origin links when a new copy is created.
These properties are hard to enforce in the concrete node type implementation. Therefore, we propose a template method `withNewChildrenInternal` that should be rewritten by the concrete classes and let the `withNewChildren` method take care of referential equality and copying:
```
override def withNewChildren(newChildren: Seq[LogicalPlan]): LogicalPlan = {
if (childrenFastEquals(children, newChildren)) {
this
} else {
CurrentOrigin.withOrigin(origin) {
val res = withNewChildrenInternal(newChildren)
res.copyTagsFrom(this)
res
}
}
}
```
With the refactoring done in a previous PR (https://github.com/apache/spark/pull/31932) most tree node types fall in one of the categories of `Leaf`, `Unary`, `Binary` or `Ternary`. These traits have a more efficient implementation for `mapChildren` and define a more specialized version of `withNewChildrenInternal` that avoids creating unnecessary lists. For example, the `mapChildren` method in `UnaryLike` is defined as follows:
```
override final def mapChildren(f: T => T): T = {
val newChild = f(child)
if (newChild fastEquals child) {
this.asInstanceOf[T]
} else {
CurrentOrigin.withOrigin(origin) {
val res = withNewChildInternal(newChild)
res.copyTagsFrom(this.asInstanceOf[T])
res
}
}
}
```
#### Results
With this PR, we have observed significant performance improvements in query compilation time, more specifically in the analysis and optimization phases. The table below shows the TPC-DS queries that had more than 25% speedup in compilation times. Biggest speedups are observed in queries with large query plans.
| Query | Speedup |
| ------------- | ------------- |
|q4 |29%|
|q9 |81%|
|q14a |31%|
|q14b |28%|
|q22 |33%|
|q33 |29%|
|q34 |25%|
|q39 |27%|
|q41 |27%|
|q44 |26%|
|q47 |28%|
|q48 |76%|
|q49 |46%|
|q56 |26%|
|q58 |43%|
|q59 |46%|
|q60 |50%|
|q65 |59%|
|q66 |46%|
|q67 |52%|
|q69 |31%|
|q70 |30%|
|q96 |26%|
|q98 |32%|
#### Binary incompatibility
Changing the `withNewChildren` in `TreeNode` breaks the binary compatibility of the code compiled against older versions of Spark because now it is expected that concrete `TreeNode` subclasses all implement the `withNewChildrenInternal` method. This is a problem, for example, when users write custom expressions. This change is the right choice, since it forces all newly added expressions to Catalyst implement it in an efficient manner and will prevent future regressions.
Please note that we have not completely removed the old implementation and renamed it to `legacyWithNewChildren`. This method will be removed in the future and for now helps the transition. There are expressions such as `UpdateFields` that have a complex way of defining children. Writing `withNewChildren` for them requires refactoring the expression. For now, these expressions use the old, slow method. In a future PR we address these expressions.
### Does this PR introduce _any_ user-facing change?
This PR does not introduce user facing changes but my break binary compatibility of the code compiled against older versions. See the binary compatibility section.
### How was this patch tested?
This PR is mainly a refactoring and passes existing tests.
Closes#32030 from dbaliafroozeh/ImprovedMapChildren.
Authored-by: Ali Afroozeh <ali.afroozeh@databricks.com>
Signed-off-by: herman <herman@databricks.com>
### What changes were proposed in this pull request?
Implement toString() and sql() methods for TRY_CAST
### Why are the changes needed?
The new expression should have a different name from `CAST` in SQL/String representation.
### Does this PR introduce _any_ user-facing change?
Yes, in the result of `explain()`, users can see try_cast if the new expression is used.
### How was this patch tested?
Unit tests.
Closes#32098 from gengliangwang/tryCastString.
Authored-by: Gengliang Wang <ltnwgl@gmail.com>
Signed-off-by: Gengliang Wang <ltnwgl@gmail.com>
### What changes were proposed in this pull request?
Change UpdateAction and InsertAction of MergeIntoTable to explicitly represent star,
### Why are the changes needed?
Currently, UpdateAction and InsertAction in the MergeIntoTable implicitly represent `update set *` and `insert *` with empty assignments. That means there is no way to differentiate between the representations of "update all columns" and "update no columns". For SQL MERGE queries, this inability does not matter because the SQL MERGE grammar that generated the MergeIntoTable plan does not allow "update no columns". However, other ways of generating the MergeIntoTable plan may not have that limitation, and may want to allow specifying "update no columns". For example, in the Delta Lake project we provide a type-safe Scala API for Merge, where it is perfectly valid to produce a Merge query with an update clause but no update assignments. Currently, we cannot use MergeIntoTable to represent this plan, thus complicating the generation, and resolution of merge query from scala API.
Side note: fixed another bug where a merge plan with star and no other expressions with unresolved attributes (e.g. all non-optional predicates are `literal(true)`), then resolution will be skipped and star wont expanded. added test for that.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existing unit tests
Closes#32067 from tdas/SPARK-34962-2.
Authored-by: Tathagata Das <tathagata.das1565@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Currently, we can't support use ordinal in CUBE/ROLLUP/GROUPING SETS,
this pr make CUBE/ROLLUP/GROUPING SETS support GROUP BY ordinal
### Why are the changes needed?
Make CUBE/ROLLUP/GROUPING SETS support GROUP BY ordinal.
Postgres SQL and TeraData support this use case.
### Does this PR introduce _any_ user-facing change?
User can use ordinal in CUBE/ROLLUP/GROUPING SETS, such as
```
-- can use ordinal in CUBE
select a, b, count(1) from data group by cube(1, 2);
-- mixed cases: can use ordinal in CUBE
select a, b, count(1) from data group by cube(1, b);
-- can use ordinal with cube
select a, b, count(1) from data group by 1, 2 with cube;
-- can use ordinal in ROLLUP
select a, b, count(1) from data group by rollup(1, 2);
-- mixed cases: can use ordinal in ROLLUP
select a, b, count(1) from data group by rollup(1, b);
-- can use ordinal with rollup
select a, b, count(1) from data group by 1, 2 with rollup;
-- can use ordinal in GROUPING SETS
select a, b, count(1) from data group by grouping sets((1), (2), (1, 2));
-- mixed cases: can use ordinal in GROUPING SETS
select a, b, count(1) from data group by grouping sets((1), (b), (a, 2));
select a, b, count(1) from data group by a, 2 grouping sets((1), (b), (a, 2));
```
### How was this patch tested?
Added UT
Closes#30145 from AngersZhuuuu/SPARK-33233.
Lead-authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Co-authored-by: angerszhu <angers.zhu@gmail.com>
Co-authored-by: AngersZhuuuu <angers.zhu@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR adds two additional checks in `CheckAnalysis` for correlated scalar subquery in Aggregate. It blocks the cases that Spark do not currently support based on the rewrite logic in `RewriteCorrelatedScalarSubquery`:
aff6c0febb/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/subquery.scala (L618-L624)
### Why are the changes needed?
It can be confusing to users when their queries pass the check analysis but cannot be executed. Also, the error messages are confusing:
#### Case 1: correlated scalar subquery in the grouping expressions but not in aggregate expressions
```sql
SELECT SUM(c2) FROM t t1 GROUP BY (SELECT SUM(c2) FROM t t2 WHERE t1.c1 = t2.c1)
```
We get this error:
```
java.lang.AssertionError: assertion failed: Expects 1 field, but got 2; something went wrong in analysis
```
because the correlated scalar subquery is not rewritten properly:
```scala
== Optimized Logical Plan ==
Aggregate [scalar-subquery#5 [(c1#6 = c1#6#93)]], [sum(c2#7) AS sum(c2)#11L]
: +- Aggregate [c1#6], [sum(c2#7) AS sum(c2)#15L, c1#6 AS c1#6#93]
: +- LocalRelation [c1#6, c2#7]
+- LocalRelation [c1#6, c2#7]
```
#### Case 2: correlated scalar subquery in the aggregate expressions but not in the grouping expressions
```sql
SELECT (SELECT SUM(c2) FROM t t2 WHERE t1.c1 = t2.c1), SUM(c2) FROM t t1 GROUP BY c1
```
We get this error:
```
java.lang.IllegalStateException: Couldn't find sum(c2)#69L in [c1#60,sum(c2#61)#64L]
```
because the transformed correlated scalar subquery output is not present in the grouping expression of the Aggregate:
```scala
== Optimized Logical Plan ==
Aggregate [c1#60], [sum(c2)#69L AS scalarsubquery(c1)#70L, sum(c2#61) AS sum(c2)#65L]
+- Project [c1#60, c2#61, sum(c2)#69L]
+- Join LeftOuter, (c1#60 = c1#60#95)
:- LocalRelation [c1#60, c2#61]
+- Aggregate [c1#60], [sum(c2#61) AS sum(c2)#69L, c1#60 AS c1#60#95]
+- LocalRelation [c1#60, c2#61]
```
### Does this PR introduce _any_ user-facing change?
Yes
### How was this patch tested?
New unit tests
Closes#32054 from allisonwang-db/spark-34946-scalar-subquery-agg.
Authored-by: allisonwang-db <66282705+allisonwang-db@users.noreply.github.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
1. Added new method `toDayTimeIntervalString()` to `IntervalUtils` which converts a day-time interval as a number of microseconds to a string in the form **"INTERVAL '[sign]days hours:minutes:secondsWithFraction' DAY TO SECOND"**.
2. Extended the `Cast` expression to support casting of `DayTimeIntervalType` to `StringType`.
### Why are the changes needed?
To conform the ANSI SQL standard which requires to support such casting.
### Does this PR introduce _any_ user-facing change?
Should not because new day-time interval has not been released yet.
### How was this patch tested?
Added new tests for casting:
```
$ build/sbt "testOnly *CastSuite*"
```
Closes#32070 from MaxGekk/cast-dt-interval-to-string.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Current trait `GroupingSet` is ambiguous, since `grouping set` in parser level means one set of a group.
Rename this to `BaseGroupingSets` since cube/rollup is syntax sugar for grouping sets.`
### Why are the changes needed?
Refactor class name
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Not need
Closes#32073 from AngersZhuuuu/SPARK-34976.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
The `explain()` method prints the arguments of tree nodes in logical/physical plans. The arguments could contain a map-type option that contains sensitive data.
We should map-type options in the output of `explain()`. Otherwise, we will see sensitive data in explain output or Spark UI.
### Why are the changes needed?
Data security.
### Does this PR introduce _any_ user-facing change?
Yes, redact the map-type options in the output of `explain()`
### How was this patch tested?
Unit tests
Closes#32066 from gengliangwang/redactOptions.
Authored-by: Gengliang Wang <ltnwgl@gmail.com>
Signed-off-by: Gengliang Wang <ltnwgl@gmail.com>
## What changes were proposed in this pull request?
This adds a new API for catalog plugins that exposes functions to Spark. The API can list and load functions. This does not include create, delete, or alter operations.
- [Design Document](https://docs.google.com/document/d/1PLBieHIlxZjmoUB0ERF-VozCRJ0xw2j3qKvUNWpWA2U/edit?usp=sharing)
There are 3 types of functions defined:
* A `ScalarFunction` that produces a value for every call
* An `AggregateFunction` that produces a value after updates for a group of rows
Functions loaded from the catalog by name as `UnboundFunction`. Once input arguments are determined `bind` is called on the unbound function to get a `BoundFunction` implementation that is one of the 3 types above. Binding can fail if the function doesn't support the input type. `BoundFunction` returns the result type produced by the function.
## How was this patch tested?
This includes a test that demonstrates the new API.
Closes#24559 from rdblue/SPARK-27658-add-function-catalog-api.
Authored-by: Ryan Blue <blue@apache.org>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This is a followup for https://github.com/apache/spark/pull/31932.
In this PR we:
- Introduce the `QuaternaryLike` trait for node types with 4 children.
- Specialize more node types
- Fix a number of style errors that were introduced in the original PR.
### Why are the changes needed?
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
This is a refactoring, passes existing tests.
Closes#32065 from dbaliafroozeh/FollowupSPARK-34906.
Authored-by: Ali Afroozeh <ali.afroozeh@databricks.com>
Signed-off-by: herman <herman@databricks.com>
### What changes were proposed in this pull request?
This PR extends the current function registry and catalog to support table-valued functions by adding a table function registry. It also refactors `range` to be a built-in function in the table function registry.
### Why are the changes needed?
Currently, Spark resolves table-valued functions very differently from the other functions. This change is to make the behavior for table and non-table functions consistent. It also allows Spark to display information about built-in table-valued functions:
Before:
```scala
scala> sql("describe function range").show(false)
+--------------------------+
|function_desc |
+--------------------------+
|Function: range not found.|
+--------------------------+
```
After:
```scala
Function: range
Class: org.apache.spark.sql.catalyst.plans.logical.Range
Usage:
range(start: Long, end: Long, step: Long, numPartitions: Int)
range(start: Long, end: Long, step: Long)
range(start: Long, end: Long)
range(end: Long)
// Extended
Function: range
Class: org.apache.spark.sql.catalyst.plans.logical.Range
Usage:
range(start: Long, end: Long, step: Long, numPartitions: Int)
range(start: Long, end: Long, step: Long)
range(start: Long, end: Long)
range(end: Long)
Extended Usage:
Examples:
> SELECT * FROM range(1);
+---+
| id|
+---+
| 0|
+---+
> SELECT * FROM range(0, 2);
+---+
|id |
+---+
|0 |
|1 |
+---+
> SELECT range(0, 4, 2);
+---+
|id |
+---+
|0 |
|2 |
+---+
Since: 2.0.0
```
### Does this PR introduce _any_ user-facing change?
Yes. User will not be able to create a function with name `range` in the default database:
Before:
```scala
scala> sql("create function range as 'range'")
res3: org.apache.spark.sql.DataFrame = []
```
After:
```
scala> sql("create function range as 'range'")
org.apache.spark.sql.catalyst.analysis.FunctionAlreadyExistsException: Function 'default.range' already exists in database 'default'
```
### How was this patch tested?
Unit test
Closes#31791 from allisonwang-db/spark-34678-table-func-registry.
Authored-by: allisonwang-db <66282705+allisonwang-db@users.noreply.github.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Changed the cost comparison function of the CBO to use the ratios of row counts and sizes in bytes.
### Why are the changes needed?
In #30965 we changed to CBO cost comparison function so it would be "symetric": `A.betterThan(B)` now implies, that `!B.betterThan(A)`.
With that we caused a performance regressions in some queries - TPCDS q19 for example.
The original cost comparison function used the ratios `relativeRows = A.rowCount / B.rowCount` and `relativeSize = A.size / B.size`. The changed function compared "absolute" cost values `costA = w*A.rowCount + (1-w)*A.size` and `costB = w*B.rowCount + (1-w)*B.size`.
Given the input from wzhfy we decided to go back to the relative values, because otherwise one (size) may overwhelm the other (rowCount). But this time we avoid adding up the ratios.
Originally `A.betterThan(B) => w*relativeRows + (1-w)*relativeSize < 1` was used. Besides being "non-symteric", this also can exhibit one overwhelming other.
For `w=0.5` If `A` size (bytes) is at least 2x larger than `B`, then no matter how many times more rows does the `B` plan have, `B` will allways be considered to be better - `0.5*2 + 0.5*0.00000000000001 > 1`.
When working with ratios, then it would be better to multiply them.
The proposed cost comparison function is: `A.betterThan(B) => relativeRows^w * relativeSize^(1-w) < 1`.
### Does this PR introduce _any_ user-facing change?
Comparison of the changed TPCDS v1.4 query execution times at sf=10:
| absolute | multiplicative | | additive |
-- | -- | -- | -- | -- | --
q12 | 145 | 137 | -5.52% | 141 | -2.76%
q13 | 264 | 271 | 2.65% | 271 | 2.65%
q17 | 4521 | 4243 | -6.15% | 4348 | -3.83%
q18 | 758 | 466 | -38.52% | 480 | -36.68%
q19 | 38503 | 2167 | -94.37% | 2176 | -94.35%
q20 | 119 | 120 | 0.84% | 126 | 5.88%
q24a | 16429 | 16838 | 2.49% | 17103 | 4.10%
q24b | 16592 | 16999 | 2.45% | 17268 | 4.07%
q25 | 3558 | 3556 | -0.06% | 3675 | 3.29%
q33 | 362 | 361 | -0.28% | 380 | 4.97%
q52 | 1020 | 1032 | 1.18% | 1052 | 3.14%
q55 | 927 | 938 | 1.19% | 961 | 3.67%
q72 | 24169 | 13377 | -44.65% | 24306 | 0.57%
q81 | 1285 | 1185 | -7.78% | 1168 | -9.11%
q91 | 324 | 336 | 3.70% | 337 | 4.01%
q98 | 126 | 129 | 2.38% | 131 | 3.97%
All times are in ms, the change is compared to the situation in the master branch (absolute).
The proposed cost function (multiplicative) significantlly improves the performance on q18, q19 and q72. The original cost function (additive) has similar improvements at q18 and q19. All other chagnes are within the error bars and I would ignore them - perhaps q81 has also improved.
### How was this patch tested?
PlanStabilitySuite
Closes#32014 from tanelk/SPARK-34922_cbo_better_cost_function.
Lead-authored-by: Tanel Kiis <tanel.kiis@gmail.com>
Co-authored-by: tanel.kiis@gmail.com <tanel.kiis@gmail.com>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
1. Added new method `toYearMonthIntervalString()` to `IntervalUtils` which converts an year-month interval as a number of month to a string in the form **"INTERVAL '[sign]yearField-monthField' YEAR TO MONTH"**.
2. Extended the `Cast` expression to support casting of `YearMonthIntervalType` to `StringType`.
### Why are the changes needed?
To conform the ANSI SQL standard which requires to support such casting.
### Does this PR introduce _any_ user-facing change?
Should not because new year-month interval has not been released yet.
### How was this patch tested?
Added new tests for casting:
```
$ build/sbt "testOnly *CastSuite*"
```
Closes#32056 from MaxGekk/cast-ym-interval-to-string.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Changes the metadata propagation framework.
Previously, most `LogicalPlan`'s propagated their `children`'s `metadataOutput`. This did not make sense in cases where the `LogicalPlan` did not even propagate their `children`'s `output`.
I set the metadata output for plans that do not propagate their `children`'s `output` to be `Nil`. Notably, `Project` and `View` no longer have metadata output.
### Why are the changes needed?
Previously, `SELECT m from (SELECT a from tb)` would output `m` if it were metadata. This did not make sense.
### Does this PR introduce _any_ user-facing change?
Yes. Now, `SELECT m from (SELECT a from tb)` will encounter an `AnalysisException`.
### How was this patch tested?
Added unit tests. I did not cover all cases, as they are fairly extensive. However, the new tests cover major cases (and an existing test already covers Join).
Closes#32017 from karenfeng/spark-34923.
Authored-by: Karen Feng <karen.feng@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
GROUP BY ... GROUPING SETS (...) is a weird SQL syntax we copied from Hive. It's not in the SQL standard or any other mainstream databases. This syntax requires users to repeat the expressions inside `GROUPING SETS (...)` after `GROUP BY`, and has a weird null semantic if `GROUP BY` contains extra expressions than `GROUPING SETS (...)`.
This PR deprecates this syntax:
1. Do not promote it in the document and only mention it as a Hive compatible sytax.
2. Simplify the code to only keep it for Hive compatibility.
### Why are the changes needed?
Deprecate a weird grammar.
### Does this PR introduce _any_ user-facing change?
No breaking change, but it removes a check to simplify the code: `GROUP BY a GROUPING SETS(a, b)` fails before and forces users to also put `b` after `GROUP BY`. Now this works just as `GROUP BY GROUPING SETS(a, b)`.
### How was this patch tested?
existing tests
Closes#32022 from cloud-fan/followup.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
https://github.com/apache/spark/pull/32015 added a way to run benchmarks much more easily in the same GitHub Actions build. This PR updates the benchmark results by using the way.
**NOTE** that looks like GitHub Actions use four types of CPU given my observations:
- Intel(R) Xeon(R) Platinum 8171M CPU 2.60GHz
- Intel(R) Xeon(R) CPU E5-2673 v4 2.30GHz
- Intel(R) Xeon(R) CPU E5-2673 v3 2.40GHz
- Intel(R) Xeon(R) Platinum 8272CL CPU 2.60GHz
Given my quick research, seems like they perform roughly similarly:
I couldn't find enough information about Intel(R) Xeon(R) Platinum 8272CL CPU 2.60GHz but the performance seems roughly similar given the numbers.
So shouldn't be a big deal especially given that this way is much easier, encourages contributors to run more and guarantee the same number of cores and same memory with the same softwares.
### Why are the changes needed?
To have a base line of the benchmarks accordingly.
### Does this PR introduce _any_ user-facing change?
No, dev-only.
### How was this patch tested?
It was generated from:
- [Run benchmarks: * (JDK 11)](https://github.com/HyukjinKwon/spark/actions/runs/713575465)
- [Run benchmarks: * (JDK 8)](https://github.com/HyukjinKwon/spark/actions/runs/713154337)
Closes#32044 from HyukjinKwon/SPARK-34950.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Use proper Java doc format for Java classes within `catalyst` module
### Why are the changes needed?
Many Java classes in `catalyst`, especially those for DataSource V2, do not have proper Java doc format. By fixing the format it helps to improve the doc's readability.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
N/A
Closes#32038 from sunchao/javadoc.
Authored-by: Chao Sun <sunchao@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
When we insert data into a partition table partition with empty DataFrame. We will call `PartitioningUtils.getPathFragment()`
then to update this partition's metadata too.
When we insert to a partition when partition value is `null`, it will throw exception like
```
[info] java.lang.NullPointerException:
[info] at scala.collection.immutable.StringOps$.length$extension(StringOps.scala:51)
[info] at scala.collection.immutable.StringOps.length(StringOps.scala:51)
[info] at scala.collection.IndexedSeqOptimized.foreach(IndexedSeqOptimized.scala:35)
[info] at scala.collection.IndexedSeqOptimized.foreach$(IndexedSeqOptimized.scala:33)
[info] at scala.collection.immutable.StringOps.foreach(StringOps.scala:33)
[info] at org.apache.spark.sql.catalyst.catalog.ExternalCatalogUtils$.escapePathName(ExternalCatalogUtils.scala:69)
[info] at org.apache.spark.sql.catalyst.catalog.ExternalCatalogUtils$.getPartitionValueString(ExternalCatalogUtils.scala:126)
[info] at org.apache.spark.sql.execution.datasources.PartitioningUtils$.$anonfun$getPathFragment$1(PartitioningUtils.scala:354)
[info] at scala.collection.TraversableLike.$anonfun$map$1(TraversableLike.scala:238)
[info] at scala.collection.Iterator.foreach(Iterator.scala:941)
[info] at scala.collection.Iterator.foreach$(Iterator.scala:941)
[info] at scala.collection.AbstractIterator.foreach(Iterator.scala:1429)
[info] at scala.collection.IterableLike.foreach(IterableLike.scala:74)
[info] at scala.collection.IterableLike.foreach$(IterableLike.scala:73)
```
`PartitioningUtils.getPathFragment()` should support `null` value too
### Why are the changes needed?
Fix bug
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Added UT
Closes#32018 from AngersZhuuuu/SPARK-34926.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
This PR introduces a new analysis rule `DeduplicateRelations`, which deduplicates any duplicate relations in a plan first and then deduplicates conflicting attributes(which resued the `dedupRight` of `ResolveReferences`).
### Why are the changes needed?
`CostBasedJoinReorder` could fail when applying on self-join, e.g.,
```scala
// test in JoinReorderSuite
test("join reorder with self-join") {
val plan = t2.join(t1, Inner, Some(nameToAttr("t1.k-1-2") === nameToAttr("t2.k-1-5")))
.select(nameToAttr("t1.v-1-10"))
.join(t2, Inner, Some(nameToAttr("t1.v-1-10") === nameToAttr("t2.k-1-5")))
// this can fail
Optimize.execute(plan.analyze)
}
```
Besides, with the new rule `DeduplicateRelations`, we'd be able to enable some optimizations, e.g., LeftSemiAnti pushdown, redundant project removal, as reflects in updated unit tests.
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
Added and updated unit tests.
Closes#32027 from Ngone51/join-reorder-3.
Lead-authored-by: yi.wu <yi.wu@databricks.com>
Co-authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR is to support nested column type in Spark ORC vectorized reader. Currently ORC vectorized reader [does not support nested column type (struct, array and map)](https://github.com/apache/spark/blob/master/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/orc/OrcFileFormat.scala#L138). We implemented nested column vectorized reader for FB-ORC in our internal fork of Spark. We are seeing performance improvement compared to non-vectorized reader when reading nested columns. In addition, this can also help improve the non-nested column performance when reading non-nested and nested columns together in one query.
Before this PR:
* `OrcColumnVector` is the implementation class for Spark's `ColumnVector` to wrap Hive's/ORC's `ColumnVector` to read `AtomicType` data.
After this PR:
* `OrcColumnVector` is an abstract class to keep interface being shared between multiple implementation class of orc column vectors, namely `OrcAtomicColumnVector` (for `AtomicType`), `OrcArrayColumnVector` (for `ArrayType`), `OrcMapColumnVector` (for `MapType`), `OrcStructColumnVector` (for `StructType`). So the original logic to read `AtomicType` data is moved from `OrcColumnVector` to `OrcAtomicColumnVector`. The abstract class of `OrcColumnVector` is needed here because of supporting nested column (i.e. nested column vectors).
* A utility method `OrcColumnVectorUtils.toOrcColumnVector` is added to create Spark's `OrcColumnVector` from Hive's/ORC's `ColumnVector`.
* A new user-facing config `spark.sql.orc.enableNestedColumnVectorizedReader` is added to control enabling/disabling vectorized reader for nested columns. The default value is false (i.e. disabling by default). For certain tables having deep nested columns, vectorized reader might take too much memory for each sub-column vectors, compared to non-vectorized reader. So providing a config here to work around OOM for query reading wide and deep nested columns if any. We plan to enable it by default on 3.3. Leave it disable in 3.2 in case for any unknown bugs.
### Why are the changes needed?
Improve query performance when reading nested columns from ORC file format.
Tested with locally adding a small benchmark in `OrcReadBenchmark.scala`. Seeing more than 1x run time improvement.
```
Running benchmark: SQL Nested Column Scan
Running case: Native ORC MR
Stopped after 2 iterations, 37850 ms
Running case: Native ORC Vectorized (Enabled Nested Column)
Stopped after 2 iterations, 15892 ms
Running case: Native ORC Vectorized (Disabled Nested Column)
Stopped after 2 iterations, 37954 ms
Running case: Hive built-in ORC
Stopped after 2 iterations, 35118 ms
Java HotSpot(TM) 64-Bit Server VM 1.8.0_181-b13 on Mac OS X 10.15.7
Intel(R) Core(TM) i9-9980HK CPU 2.40GHz
SQL Nested Column Scan: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------------
Native ORC MR 18706 18925 310 0.1 17839.6 1.0X
Native ORC Vectorized (Enabled Nested Column) 7625 7946 455 0.1 7271.6 2.5X
Native ORC Vectorized (Disabled Nested Column) 18415 18977 796 0.1 17561.5 1.0X
Hive built-in ORC 17469 17559 127 0.1 16660.1 1.1X
```
Benchmark:
```
nestedColumnScanBenchmark(1024 * 1024)
def nestedColumnScanBenchmark(values: Int): Unit = {
val benchmark = new Benchmark(s"SQL Nested Column Scan", values, output = output)
withTempPath { dir =>
withTempTable("t1", "nativeOrcTable", "hiveOrcTable") {
import spark.implicits._
spark.range(values).map(_ => Random.nextLong).map { x =>
val arrayOfStructColumn = (0 until 5).map(i => (x + i, s"$x" * 5))
val mapOfStructColumn = Map(
s"$x" -> (x * 0.1, (x, s"$x" * 100)),
(s"$x" * 2) -> (x * 0.2, (x, s"$x" * 200)),
(s"$x" * 3) -> (x * 0.3, (x, s"$x" * 300)))
(arrayOfStructColumn, mapOfStructColumn)
}.toDF("col1", "col2")
.createOrReplaceTempView("t1")
prepareTable(dir, spark.sql(s"SELECT * FROM t1"))
benchmark.addCase("Native ORC MR") { _ =>
withSQLConf(SQLConf.ORC_VECTORIZED_READER_ENABLED.key -> "false") {
spark.sql("SELECT SUM(SIZE(col1)), SUM(SIZE(col2)) FROM nativeOrcTable").noop()
}
}
benchmark.addCase("Native ORC Vectorized (Enabled Nested Column)") { _ =>
spark.sql("SELECT SUM(SIZE(col1)), SUM(SIZE(col2)) FROM nativeOrcTable").noop()
}
benchmark.addCase("Native ORC Vectorized (Disabled Nested Column)") { _ =>
withSQLConf(SQLConf.ORC_VECTORIZED_READER_NESTED_COLUMN_ENABLED.key -> "false") {
spark.sql("SELECT SUM(SIZE(col1)), SUM(SIZE(col2)) FROM nativeOrcTable").noop()
}
}
benchmark.addCase("Hive built-in ORC") { _ =>
spark.sql("SELECT SUM(SIZE(col1)), SUM(SIZE(col2)) FROM hiveOrcTable").noop()
}
benchmark.run()
}
}
}
```
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Added one simple test in `OrcSourceSuite.scala` to verify correctness.
Definitely need more unit tests and add benchmark here, but I want to first collect feedback before crafting more tests.
Closes#31958 from c21/orc-vector.
Authored-by: Cheng Su <chengsu@fb.com>
Signed-off-by: Liang-Chi Hsieh <viirya@gmail.com>
### What changes were proposed in this pull request?
Modify the `SubtractTimestamps` expression to return values of `DayTimeIntervalType` when `spark.sql.legacy.interval.enabled` is set to `false` (which is the default).
### Why are the changes needed?
To conform to the ANSI SQL standard which requires ANSI intervals as the result of timestamps subtraction, see
<img width="656" alt="Screenshot 2021-03-29 at 19 09 34" src="https://user-images.githubusercontent.com/1580697/112866455-7e2f0d00-90c2-11eb-96e6-3feb7eea7e09.png">
### Does this PR introduce _any_ user-facing change?
Yes.
### How was this patch tested?
By running new tests:
```
$ build/sbt "test:testOnly *DateTimeUtilsSuite"
$ build/sbt "test:testOnly *DateExpressionsSuite"
$ build/sbt "test:testOnly *ColumnExpressionSuite"
```
and some tests from `SQLQueryTestSuite`:
```
$ build/sbt "sql/testOnly *SQLQueryTestSuite -- -z timestamp.sql"
$ build/sbt "sql/testOnly *SQLQueryTestSuite -- -z datetime.sql"
$ build/sbt "sql/testOnly *SQLQueryTestSuite -- -z interval.sql"
```
Closes#32016 from MaxGekk/subtract-timestamps-to-intervals.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Change partitioning to `SinglePartition`.
### Why are the changes needed?
For node `Repartition` and `RepartitionByExpression`, if partition number is 1 we can use `SinglePartition` instead of other `Partitioning`.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Add test
Closes#32012 from ulysses-you/SPARK-34919.
Authored-by: ulysses-you <ulyssesyou18@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR fixes JDK 11 compilation failed:
```
/home/runner/work/spark/spark/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/TryCast.scala:35: error: annotation argument needs to be a constant; found: "_FUNC_(expr AS type) - Casts the value `expr` to the target data type `type`. ".+("This expression is identical to CAST with configuration `spark.sql.ansi.enabled` as ").+("true, except it returns NULL instead of raising an error. Note that the behavior of this ").+("expression doesn\'t depend on configuration `spark.sql.ansi.enabled`.")
"true, except it returns NULL instead of raising an error. Note that the behavior of this " +
```
For whatever reason, it doesn't know that the string is actually a constant. This PR simply switches it to multi-line style (which is actually more correct).
Reference:
bd0990e3e8/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/ApproximatePercentile.scala (L53-L57)
### Why are the changes needed?
To recover the build.
### Does this PR introduce _any_ user-facing change?
No, dev-only.
### How was this patch tested?
CI in this PR
Closes#32019 from HyukjinKwon/SPARK-34881.
Lead-authored-by: Hyukjin Kwon <gurwls223@apache.org>
Co-authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: Gengliang Wang <ltnwgl@gmail.com>
### What changes were proposed in this pull request?
Replaced the `agg(if (('gid = 1)) 'cat1 else null)` pattern in `RewriteDistinctAggregates` with `agg('cat1) FILTER (WHERE 'gid = 1)`
### Why are the changes needed?
For aggregate functions, that do not ignore NULL values (`First`, `Last` or `UDAF`s) the current approach can return wrong results.
In the added UT there are no nulls in the input `testData`. The query returned `Row(0, 1, 0, 51, 100)` before this PR.
### Does this PR introduce _any_ user-facing change?
Bugfix
### How was this patch tested?
UT
Closes#31983 from tanelk/SPARK-34882_distinct_agg_filter.
Lead-authored-by: Tanel Kiis <tanel.kiis@gmail.com>
Co-authored-by: tanel.kiis@gmail.com <tanel.kiis@gmail.com>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
Add a new SQL function `try_cast`.
`try_cast` is identical to `AnsiCast` (or `Cast` when `spark.sql.ansi.enabled` is true), except it returns NULL instead of raising an error.
This expression has one major difference from `cast` with `spark.sql.ansi.enabled` as true: when the source value can't be stored in the target integral(Byte/Short/Int/Long) type, `try_cast` returns null instead of returning the low order bytes of the source value.
Note that the result of `try_cast` is not affected by the configuration `spark.sql.ansi.enabled`.
This is learned from Google BigQuery and Snowflake:
https://docs.snowflake.com/en/sql-reference/functions/try_cast.htmlhttps://cloud.google.com/bigquery/docs/reference/standard-sql/functions-and-operators#safe_casting
### Why are the changes needed?
This is an useful for the following scenarios:
1. When ANSI mode is on, users can choose `try_cast` an alternative way to run SQL without errors for certain operations.
2. When ANSI mode is off, users can use `try_cast` to get a more reasonable result for casting a value to an integral type: when an overflow error happens, `try_cast` returns null while `cast` returns the low order bytes of the source value.
### Does this PR introduce _any_ user-facing change?
Yes, adding a new function `try_cast`
### How was this patch tested?
Unit tests.
Closes#31982 from gengliangwang/tryCast.
Authored-by: Gengliang Wang <ltnwgl@gmail.com>
Signed-off-by: Gengliang Wang <ltnwgl@gmail.com>
### What changes were proposed in this pull request?
This PR introduces a new analysis rule `DeduplicateRelations`, which deduplicates any duplicate relations in a plan first and then deduplicates conflicting attributes(which resued the `dedupRight` of `ResolveReferences`).
### Why are the changes needed?
`CostBasedJoinReorder` could fail when applying on self-join, e.g.,
```scala
// test in JoinReorderSuite
test("join reorder with self-join") {
val plan = t2.join(t1, Inner, Some(nameToAttr("t1.k-1-2") === nameToAttr("t2.k-1-5")))
.select(nameToAttr("t1.v-1-10"))
.join(t2, Inner, Some(nameToAttr("t1.v-1-10") === nameToAttr("t2.k-1-5")))
// this can fail
Optimize.execute(plan.analyze)
}
```
Besides, with the new rule `DeduplicateRelations`, we'd be able to enable some optimizations, e.g., LeftSemiAnti pushdown, redundant project removal, as reflects in updated unit tests.
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
Added and updated unit tests.
Closes#31470 from Ngone51/join-reorder.
Lead-authored-by: yi.wu <yi.wu@databricks.com>
Co-authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
1. Add the SQL config `spark.sql.legacy.interval.enabled` which will control when Spark SQL should use `CalendarIntervalType` instead of ANSI intervals.
2. Modify the `SubtractDates` expression to return values of `DayTimeIntervalType` when `spark.sql.legacy.interval.enabled` is set to `false` (which is the default).
### Why are the changes needed?
To conform to the ANSI SQL standard which requires ANSI intervals as the result of dates subtraction, see
<img width="656" alt="Screenshot 2021-03-29 at 19 09 34" src="https://user-images.githubusercontent.com/1580697/112866455-7e2f0d00-90c2-11eb-96e6-3feb7eea7e09.png">
### Does this PR introduce _any_ user-facing change?
Yes.
### How was this patch tested?
By running new tests:
```
$ build/sbt "test:testOnly *DateExpressionsSuite"
$ build/sbt "test:testOnly *ColumnExpressionSuite"
```
and some tests from `SQLQueryTestSuite`:
```
$ build/sbt "sql/testOnly *SQLQueryTestSuite -- -z date.sql"
$ build/sbt "sql/testOnly *SQLQueryTestSuite -- -z datetime.sql"
$ build/sbt "sql/testOnly *SQLQueryTestSuite -- -z interval.sql"
```
Closes#31996 from MaxGekk/subtract-dates-to-intervals.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Use `java.lang.Long.divideUnsigned()` to do integer division in `NumberConverter` to avoid a bug in `unsignedLongDiv` that produced invalid results.
### Why are the changes needed?
The previous results are incorrect, the result of the below query should be 45012021522523134134555
```
scala> spark.sql("select conv('-10', 11, 7)").show(20, 150)
+-----------------------+
| conv(-10, 11, 7)|
+-----------------------+
|4501202152252313413456|
+-----------------------+
scala> spark.sql("select hex(conv('-10', 11, 7))").show(20, 150)
+----------------------------------------------+
| hex(conv(-10, 11, 7))|
+----------------------------------------------+
|3435303132303231353232353233313334313334353600|
+----------------------------------------------+
```
### Does this PR introduce _any_ user-facing change?
`conv()` will produce different results because the bug is fixed.
### How was this patch tested?
Added a simple unit test.
Closes#32006 from timarmstrong/conv-unsigned.
Authored-by: Tim Armstrong <tim.armstrong@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
After https://github.com/apache/spark/pull/31954/, Array type is allowed to be cast as String type. So the customized conversion failure message branch from AnsiCast.typeCheckFailureMessage won't be reached anymore.
This PR is to remove the dead code.
### Why are the changes needed?
Code clean up.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Just removing dead code.
Closes#32004 from gengliangwang/SPARK-34856-followup.
Authored-by: Gengliang Wang <ltnwgl@gmail.com>
Signed-off-by: Sean Owen <srowen@gmail.com>
### What changes were proposed in this pull request?
Spark query plan node hierarchy has specialized traits (or abstract classes) for handling nodes with fixed number of children, for example `UnaryExpression`, `UnaryNode` and `UnaryExec` for representing an expression, a logical plan and a physical plan with only one child, respectively. This PR refactors the `TreeNode` hierarchy by extracting the children handling functionality into the following traits. `UnaryExpression` and other similar classes now extend the corresponding new trait:
```
trait LeafLike[T <: TreeNode[T]] { self: TreeNode[T] =>
override final def children: Seq[T] = Nil
}
trait UnaryLike[T <: TreeNode[T]] { self: TreeNode[T] =>
def child: T
transient override final lazy val children: Seq[T] = child :: Nil
}
trait BinaryLike[T <: TreeNode[T]] { self: TreeNode[T] =>
def left: T
def right: T
transient override final lazy val children: Seq[T] = left :: right :: Nil
}
trait TernaryLike[T <: TreeNode[T]] { self: TreeNode[T] =>
def first: T
def second: T
def third: T
transient override final lazy val children: Seq[T] = first :: second :: third :: Nil
}
```
This refactoring, which is part of a bigger effort to make tree transformations in Spark more efficient, has two benefits:
- It moves the children handling methods to a single place, instead of being spread in specific subclasses, which will help the future optimizations for tree traversals.
- It allows to mix in these traits with some concrete node types that could not extend the previous classes. For example, expressions with one child that extend `AggregateFunction` cannot extend `UnaryExpression` as `AggregateFunction` defines the `foldable` method final while `UnaryExpression` defines it as non final. With the new traits, we can directly extend the concrete class from `UnaryLike` in these cases. Classes with more specific child handling will make tree traversal methods faster.
In this PR we have also updated many concrete node types to extend these traits to benefit from more specific child handling.
### Why are the changes needed?
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
This is a refactoring, passes existing tests.
Closes#31932 from dbaliafroozeh/FactorOutChildHandlnigIntoSeparateTraits.
Authored-by: Ali Afroozeh <ali.afroozeh@databricks.com>
Signed-off-by: herman <herman@databricks.com>
### What changes were proposed in this pull request?
As discussed in
https://github.com/apache/spark/pull/30145#discussion_r514728642https://github.com/apache/spark/pull/30145#discussion_r514734648
We need to rewrite current Grouping Analytics grammar to support as flexible as Postgres SQL to support subsequent development.
In postgres sql, it support
```
select a, b, c, count(1) from t group by cube (a, b, c);
select a, b, c, count(1) from t group by cube(a, b, c);
select a, b, c, count(1) from t group by cube (a, b, c, (a, b), (a, b, c));
select a, b, c, count(1) from t group by rollup(a, b, c);
select a, b, c, count(1) from t group by rollup (a, b, c);
select a, b, c, count(1) from t group by rollup (a, b, c, (a, b), (a, b, c));
```
In this pr, we have done three things as below, and we will split it to different pr:
- Refactor CUBE/ROLLUP (regarding them as ANTLR tokens in a parser)
- Refactor GROUPING SETS (the logical node -> a new expr)
- Support new syntax for CUBE/ROLLUP (e.g., GROUP BY CUBE ((a, b), (a, c)))
### Why are the changes needed?
Rewrite current Grouping Analytics grammar to support as flexible as Postgres SQL to support subsequent development.
### Does this PR introduce _any_ user-facing change?
User can write Grouping Analytics grammar as flexible as Postgres SQL to support subsequent development.
### How was this patch tested?
Added UT
Closes#30212 from AngersZhuuuu/refact-grouping-analytics.
Lead-authored-by: angerszhu <angers.zhu@gmail.com>
Co-authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Co-authored-by: AngersZhuuuu <angers.zhu@gmail.com>
Co-authored-by: Takeshi Yamamuro <yamamuro@apache.org>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
As a followup for #29818, document caveats of using the Arrow selfDestruct option in toPandas, which include:
- toPandas() may be slower;
- the resulting dataframe may not support some Pandas operations due to immutable backing arrays.
### Why are the changes needed?
This will hopefully reduce user confusion as with SPARK-34463.
### Does this PR introduce _any_ user-facing change?
Yes - documentation is updated and a config setting description is updated to clearly indicate the config is experimental.
### How was this patch tested?
This is a documentation-only change.
Closes#31738 from lidavidm/spark-34463.
Authored-by: David Li <li.davidm96@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Some `spark-submit` commands used to run benchmarks in the user's guide is wrong, we can't use these commands to run benchmarks successful.
So the major changes of this pr is correct these wrong commands, for example, run a benchmark which inherits from `SqlBasedBenchmark`, we must specify `--jars <spark core test jar>,<spark catalyst test jar>` because `SqlBasedBenchmark` based benchmark extends `BenchmarkBase(defined in spark core test jar)` and `SQLHelper(defined in spark catalyst test jar)`.
Another change of this pr is removed the `scalatest Assertions` dependency of Benchmarks because `scalatest-*.jar` are not in the distribution package, it will be troublesome to use.
### Why are the changes needed?
Make sure benchmarks can run using spark-submit cmd described in the guide
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Use the corrected `spark-submit` commands to run benchmarks successfully.
Closes#31995 from LuciferYang/fix-benchmark-guide.
Authored-by: yangjie01 <yangjie01@baidu.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
There is a `Project` between `LocalLimit` and `Join` if `Join`'s output do not match the `LocalLimit`'s output. This pr add support push down limit through this case. For example:
```scala
spark.sql("create table t1(a int, b int, c int) using parquet")
spark.sql("create table t2(x int, y int, z int) using parquet")
spark.sql("select a from t1 left join t2 on a = x and b = y limit 5").explain("extended")
```
```
== Optimized Logical Plan ==
GlobalLimit 5
+- LocalLimit 5
+- Project [a#0]
+- Join LeftOuter, ((a#0 = x#3) AND (b#1 = y#4))
:- Project [a#0, b#1]
: +- Relation default.t1[a#0,b#1,c#2] parquet
+- Project [x#3, y#4]
+- Filter (isnotnull(x#3) AND isnotnull(y#4))
+- Relation default.t2[x#3,y#4,z#5] parquet
```
After this pr:
```
== Optimized Logical Plan ==
GlobalLimit 5
+- LocalLimit 5
+- Project [a#0]
+- Join LeftOuter, ((a#0 = x#3) AND (b#1 = y#4))
:- LocalLimit 5
: +- Project [a#0, b#1]
: +- Relation default.t1[a#0,b#1,c#2] parquet
+- Project [x#3, y#4]
+- Filter (isnotnull(x#3) AND isnotnull(y#4))
+- Relation default.t2[x#3,y#4,z#5] parquet
```
### Why are the changes needed?
Improve limit push down to improve query performance.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Unit test.
Closes#31739 from wangyum/SPARK-34622.
Authored-by: Yuming Wang <yumwang@ebay.com>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
This PR proposes to extend the functionality of requirement for distribution and ordering on V2 write to specify the number of partitioning on repartition, so that data source is able to control the parallelism and determine the data distribution per partition in prior.
The partitioning with static number is optional, and by default disabled via default method, so only implementations required to restrict the number of partition statically need to override the method and provide the number.
Note that we don't support static number of partitions with unspecified distribution for this PR, as we haven't found the real use cases, and for hypothetical case the static number isn't good enough. Javadoc clearly describes the limitation.
### Why are the changes needed?
The use case comes from feature parity with DSv1.
I have state data source which enables the state in SS to be rewritten, which enables repartitioning, schema evolution, etc via batch query. The writer requires hash partitioning against group key, with the "desired number of partitions", which is same as what Spark does read and write against state.
This is now implemented as DSv1, and the requirement is simply done by calling repartition with the "desired number".
```
val fullPathsForKeyColumns = keySchema.map(key => new Column(s"key.${key.name}"))
data
.repartition(newPartitions, fullPathsForKeyColumns: _*)
.queryExecution
.toRdd
.foreachPartition(
writeFn(resolvedCpLocation, version, operatorId, storeName, keySchema, valueSchema,
storeConf, hadoopConfBroadcast, queryId))
```
Thanks to SPARK-34026, it's now possible to require the hash partitioning, but still not able to require the number of partitions. This PR will enable to let data source require the number of partitions.
### Does this PR introduce _any_ user-facing change?
Yes, but only for data source implementors. Even for them, this is no breaking change as default method is added.
### How was this patch tested?
Added UTs.
Closes#31355 from HeartSaVioR/SPARK-34255.
Lead-authored-by: Jungtaek Lim <kabhwan.opensource@gmail.com>
Co-authored-by: Jungtaek Lim (HeartSaVioR) <kabhwan.opensource@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR fixes an issue that `quoteIfNeeded` quotes a name only if it contains `.` or ``` ` ```.
This method should quote it if it contains non-word characters.
### Why are the changes needed?
It's a potential bug.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
New test.
Closes#31964 from sarutak/fix-quoteIfNeeded.
Authored-by: Kousuke Saruta <sarutak@oss.nttdata.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
LikeSimplification should handle NULL.
UT will failed before this pr
```
test("SPARK-34814: LikeSimplification should handle NULL") {
withSQLConf(SQLConf.OPTIMIZER_EXCLUDED_RULES.key ->
ConstantFolding.getClass.getName.stripSuffix("$")) {
checkEvaluation(Literal.create("foo", StringType)
.likeAll("%foo%", Literal.create(null, StringType)), null)
}
}
[info] - test *** FAILED *** (2 seconds, 443 milliseconds)
[info] java.lang.NullPointerException:
[info] at org.apache.spark.sql.catalyst.optimizer.LikeSimplification$.$anonfun$simplifyMultiLike$1(expressions.scala:697)
[info] at scala.collection.TraversableLike.$anonfun$map$1(TraversableLike.scala:238)
[info] at scala.collection.mutable.ResizableArray.foreach(ResizableArray.scala:62)
[info] at scala.collection.mutable.ResizableArray.foreach$(ResizableArray.scala:55)
[info] at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:49)
[info] at scala.collection.TraversableLike.map(TraversableLike.scala:238)
[info] at scala.collection.TraversableLike.map$(TraversableLike.scala:231)
[info] at scala.collection.AbstractTraversable.map(Traversable.scala:108)
[info] at org.apache.spark.sql.catalyst.optimizer.LikeSimplification$.org$apache$spark$sql$catalyst$optimizer$LikeSimplification$$simplifyMultiLike(expressions.scala:697)
[info] at org.apache.spark.sql.catalyst.optimizer.LikeSimplification$$anonfun$apply$9.applyOrElse(expressions.scala:722)
[info] at org.apache.spark.sql.catalyst.optimizer.LikeSimplification$$anonfun$apply$9.applyOrElse(expressions.scala:714)
[info] at org.apache.spark.sql.catalyst.trees.TreeNode.$anonfun$transformDown$1(TreeNode.scala:316)
[info] at org.apache.spark.sql.catalyst.trees.CurrentOrigin$.withOrigin(TreeNode.scala:72)
[info] at org.apache.spark.sql.catalyst.trees.TreeNode.transformDown(TreeNode.scala:316)
[info] at org.apache.spark.sql.catalyst.trees.TreeNode.$anonfun$transformDown$3(TreeNode.scala:321)
[info] at org.apache.spark.sql.catalyst.trees.TreeNode.$anonfun$mapChildren$1(TreeNode.scala:406)
[info] at org.apache.spark.sql.catalyst.trees.TreeNode.mapProductIterator(TreeNode.scala:242)
[info] at org.apache.spark.sql.catalyst.trees.TreeNode.mapChildren(TreeNode.scala:404)
[info] at org.apache.spark.sql.catalyst.trees.TreeNode.mapChildren(TreeNode.scala:357)
[info] at org.apache.spark.sql.catalyst.trees.TreeNode.transformDown(TreeNode.scala:321)
[info] at org.apache.spark.sql.catalyst.plans.QueryPlan.$anonfun$transformExpressionsDown$1(QueryPlan.scala:94)
[info] at org.apache.spark.sql.catalyst.plans.QueryPlan.$anonfun$mapExpressions$1(QueryPlan.scala:116)
[info] at org.apache.spark.sql.catalyst.trees.CurrentOrigin$.withOrigin(TreeNode.scala:72)
```
### Why are the changes needed?
Fix bug
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Added UT
Closes#31976 from AngersZhuuuu/SPARK-34814.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Filled the `defaultResult` field on non-nullable aggregates
### Why are the changes needed?
The `defaultResult` defaults to `None` and in some situations (like correlated scalar subqueries) it is used for the value of the aggregation.
The UT result before the fix:
```
-- !query
SELECT t1a,
(SELECT count(t2d) FROM t2 WHERE t2a = t1a) count_t2,
(SELECT count_if(t2d > 0) FROM t2 WHERE t2a = t1a) count_if_t2,
(SELECT approx_count_distinct(t2d) FROM t2 WHERE t2a = t1a) approx_count_distinct_t2,
(SELECT collect_list(t2d) FROM t2 WHERE t2a = t1a) collect_list_t2,
(SELECT collect_set(t2d) FROM t2 WHERE t2a = t1a) collect_set_t2,
(SELECT hex(count_min_sketch(t2d, 0.5d, 0.5d, 1)) FROM t2 WHERE t2a = t1a) collect_set_t2
FROM t1
-- !query schema
struct<t1a:string,count_t2:bigint,count_if_t2:bigint,approx_count_distinct_t2:bigint,collect_list_t2:array<bigint>,collect_set_t2:array<bigint>,collect_set_t2:string>
-- !query output
val1a 0 0 NULL NULL NULL NULL
val1a 0 0 NULL NULL NULL NULL
val1a 0 0 NULL NULL NULL NULL
val1a 0 0 NULL NULL NULL NULL
val1b 6 6 3 [19,119,319,19,19,19] [19,119,319] 0000000100000000000000060000000100000004000000005D8D6AB90000000000000000000000000000000400000000000000010000000000000001
val1c 2 2 2 [219,19] [219,19] 0000000100000000000000020000000100000004000000005D8D6AB90000000000000000000000000000000100000000000000000000000000000001
val1d 0 0 NULL NULL NULL NULL
val1d 0 0 NULL NULL NULL NULL
val1d 0 0 NULL NULL NULL NULL
val1e 1 1 1 [19] [19] 0000000100000000000000010000000100000004000000005D8D6AB90000000000000000000000000000000100000000000000000000000000000000
val1e 1 1 1 [19] [19] 0000000100000000000000010000000100000004000000005D8D6AB90000000000000000000000000000000100000000000000000000000000000000
val1e 1 1 1 [19] [19] 0000000100000000000000010000000100000004000000005D8D6AB90000000000000000000000000000000100000000000000000000000000000000
```
### Does this PR introduce _any_ user-facing change?
Bugfix
### How was this patch tested?
UT
Closes#31973 from tanelk/SPARK-34876_non_nullable_agg_subquery.
Authored-by: Tanel Kiis <tanel.kiis@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR fixes a correctness issue with higher order functions. The results of function expressions needs to be copied in some higher order functions as such an expression can return with internal buffers and higher order functions can call multiple times the expression.
The issue was discovered with typed `ScalaUDF`s after https://github.com/apache/spark/pull/28979.
### Why are the changes needed?
To fix a bug.
### Does this PR introduce _any_ user-facing change?
Yes, some queries return the right results again.
### How was this patch tested?
Added new UT.
Closes#31955 from peter-toth/SPARK-34829-fix-scalaudf-resultconversion.
Authored-by: Peter Toth <peter.toth@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Fix code format in `SQLConf` and comment in `PartitionPruning`.
### Why are the changes needed?
Make code more readable.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
N/A
Closes#31969 from wangyum/SPARK-32855-2.
Authored-by: Yuming Wang <yumwang@ebay.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Push ANSI interval binary expressions into into (if / case) branches
### Why are the changes needed?
Support more binary expression to push into if/else and casewhen
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Added UT
Closes#31978 from AngersZhuuuu/SPARK-34841.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Improve error message for casting cause overflow error. We should use DataType's catalogString.
### Why are the changes needed?
Improve error message
### Does this PR introduce _any_ user-facing change?
For example:
```
set spark.sql.ansi.enabled=true;
select tinyint(128) * tinyint(2);
```
Error message before this pr:
```
Casting 128 to scala.Byte$ causes overflow
```
After this pr:
```
Casting 128 to tinyint causes overflow
```
### How was this patch tested?
Added UT
Closes#31971 from AngersZhuuuu/SPARK-34744.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Kent Yao <yao@apache.org>
### What changes were proposed in this pull request?
1. Add new expression `DivideDTInterval` which multiplies a `DayTimeIntervalType` expression by a `NumericType` expression including ByteType, ShortType, IntegerType, LongType, FloatType, DoubleType, DecimalType.
2. Extend binary arithmetic rules to support `day-time interval / numeric`.
### Why are the changes needed?
To conform the ANSI SQL standard which requires such operation over day-time intervals:
<img width="656" alt="Screenshot 2021-03-25 at 18 44 58" src="https://user-images.githubusercontent.com/1580697/112501559-68f07080-8d9a-11eb-8781-66e6631bb7ef.png">
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
By running new tests:
```
$ build/sbt "test:testOnly *IntervalExpressionsSuite"
$ build/sbt "test:testOnly *ColumnExpressionSuite"
```
Closes#31972 from MaxGekk/div-dt-interval-by-num.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
1. Add new expression `DivideYMInterval` which multiplies a `YearMonthIntervalType` expression by a `NumericType` expression including ByteType, ShortType, IntegerType, LongType, FloatType, DoubleType, DecimalType.
2. Extend binary arithmetic rules to support `year-month interval / numeric`.
### Why are the changes needed?
To conform the ANSI SQL standard which requires such operation over year-month intervals:
<img width="656" alt="Screenshot 2021-03-25 at 18 44 58" src="https://user-images.githubusercontent.com/1580697/112501559-68f07080-8d9a-11eb-8781-66e6631bb7ef.png">
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
By running new tests:
```
$ build/sbt "test:testOnly *IntervalExpressionsSuite"
$ build/sbt "test:testOnly *ColumnExpressionSuite"
```
Closes#31961 from MaxGekk/div-ym-interval-by-num.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This pr improve the cost model in `pruningHasBenefit` for filtering side can not build broadcast by join type:
1. The filtering side must be small enough to build broadcast by size.
2. The estimated size of the pruning side must be big enough: `estimatePruningSideSize * spark.sql.optimizer.dynamicPartitionPruning.pruningSideExtraFilterRatio > overhead`.
### Why are the changes needed?
Improve query performance for these cases.
This a real case from cluster. Left join and left size very small and right side can build DPP:
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Unit test.
Closes#29726 from wangyum/SPARK-32855.
Authored-by: Yuming Wang <yumwang@ebay.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
A companion PR for SPARK-34817, when we handle the unsigned int(<=32) logical types. In this PR, we map the unsigned int64 to decimal(20, 0) for better compatibility.
### Why are the changes needed?
Spark won't have unsigned types, but spark should be able to read existing parquet files written by other systems that support unsigned types for better compatibility.
### Does this PR introduce _any_ user-facing change?
yes, we can read parquet uint64 now
### How was this patch tested?
new unit tests
Closes#31960 from yaooqinn/SPARK-34786-2.
Authored-by: Kent Yao <yao@apache.org>
Signed-off-by: Kent Yao <yao@apache.org>
### What changes were proposed in this pull request?
Move the checkpoint location resolving into the rule ResolveWriteToStream, which is added in SPARK-34748.
### Why are the changes needed?
After SPARK-34748, we have a rule ResolveWriteToStream for the analysis logic for the resolving logic of stream write plans. Based on it, we can further move the checkpoint location resolving work in the rule. Then, all the checkpoint resolving logic was done in the analyzer.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Existing UT.
Closes#31963 from xuanyuanking/SPARK-34871.
Authored-by: Yuanjian Li <yuanjian.li@databricks.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This is a follow-up of https://github.com/apache/spark/pull/31940 . This PR generalizes the matching of attributes and outer references, so that outer references are handled everywhere.
Note that, currently correlated subquery has a lot of limitations in Spark, and the newly covered cases are not possible to happen. So this PR is a code refactor.
### Why are the changes needed?
code cleanup
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
existing tests
Closes#31959 from cloud-fan/follow.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
Allow casting complex types as string type in ANSI mode.
### Why are the changes needed?
Currently, complex types are not allowed to cast as string type. This breaks the DataFrame.show() API. E.g
```
scala> sql(“select array(1, 2, 2)“).show(false)
org.apache.spark.sql.AnalysisException: cannot resolve ‘CAST(`array(1, 2, 2)` AS STRING)’ due to data type mismatch:
cannot cast array<int> to string with ANSI mode on.
```
We should allow the conversion as the extension of the ANSI SQL standard, so that the DataFrame.show() still work in ANSI mode.
### Does this PR introduce _any_ user-facing change?
Yes, casting complex types as string type is now allowed in ANSI mode.
### How was this patch tested?
Unit tests.
Closes#31954 from gengliangwang/fixExplicitCast.
Authored-by: Gengliang Wang <ltnwgl@gmail.com>
Signed-off-by: Gengliang Wang <ltnwgl@gmail.com>
### What changes were proposed in this pull request?
This PR groups exception messages in `execution/datasources/v2`.
### Why are the changes needed?
It will largely help with standardization of error messages and its maintenance.
### Does this PR introduce _any_ user-facing change?
No. Error messages remain unchanged.
### How was this patch tested?
No new tests - pass all original tests to make sure it doesn't break any existing behavior.
Closes#31619 from karenfeng/spark-33600.
Authored-by: Karen Feng <karen.feng@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Removed the custom toString implementation of AtLeastNNoneNulls.
### Why are the changes needed?
It shows up wrong in the explain plan. The name of the function is wrong and the actual value of the first argument is not shown. Both of these would make it easier to understand the plan.
```
(12) Filter
Input [3]: [c1#2410L, c2#2419, c3#2422]
Condition : AtLeastNNulls(n, c1#2410L)
```
### Does this PR introduce _any_ user-facing change?
Only the explain plan changes if this function is used.
### How was this patch tested?
Added a simple unit test to make sure that the toString output is correct.
Closes#31956 from timarmstrong/atleastnnonnulls.
Authored-by: Tim Armstrong <tim.armstrong@databricks.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
1. Add new expression `MultiplyDTInterval` which multiplies a `DayTimeIntervalType` expression by a `NumericType` expression including ByteType, ShortType, IntegerType, LongType, FloatType, DoubleType, DecimalType.
2. Extend binary arithmetic rules to support `numeric * day-time interval` and `day-time interval * numeric`.
3. Invoke `DoubleMath.roundToInt` in `double/float * year-month interval`.
### Why are the changes needed?
To conform the ANSI SQL standard which requires such operation over day-time intervals:
<img width="667" alt="Screenshot 2021-03-22 at 16 33 16" src="https://user-images.githubusercontent.com/1580697/111997810-77d1eb80-8b2c-11eb-951d-e43911d9c5db.png">
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
By running new tests:
```
$ build/sbt "test:testOnly *IntervalExpressionsSuite"
$ build/sbt "test:testOnly *ColumnExpressionSuite"
```
Closes#31951 from MaxGekk/mul-day-time-interval.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
This PR proposes to remove re-analyzing the already analyzed plan for `CreateViewCommand` as discussed https://github.com/apache/spark/pull/31273/files#r581592786.
### Why are the changes needed?
No need to analyze the plan if it's already analyzed.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Existing tests should cover this.
Closes#31933 from imback82/remove_analyzed_from_create_temp_view.
Authored-by: Terry Kim <yuminkim@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR intends to fix the bug that does not apply right-padding for char types inside correlated subquries.
For example, a query below returns nothing in master, but a correct result is `c`.
```
scala> sql(s"CREATE TABLE t1(v VARCHAR(3), c CHAR(5)) USING parquet")
scala> sql(s"CREATE TABLE t2(v VARCHAR(5), c CHAR(7)) USING parquet")
scala> sql("INSERT INTO t1 VALUES ('c', 'b')")
scala> sql("INSERT INTO t2 VALUES ('a', 'b')")
scala> val df = sql("""
|SELECT v FROM t1
|WHERE 'a' IN (SELECT v FROM t2 WHERE t2.c = t1.c )""".stripMargin)
scala> df.show()
+---+
| v|
+---+
+---+
```
This is because `ApplyCharTypePadding` does not handle the case above to apply right-padding into `'abc'`. This PR modifies the code in `ApplyCharTypePadding` for handling it correctly.
```
// Before this PR:
scala> df.explain(true)
== Analyzed Logical Plan ==
v: string
Project [v#13]
+- Filter a IN (list#12 [c#14])
: +- Project [v#15]
: +- Filter (c#16 = outer(c#14))
: +- SubqueryAlias spark_catalog.default.t2
: +- Relation default.t2[v#15,c#16] parquet
+- SubqueryAlias spark_catalog.default.t1
+- Relation default.t1[v#13,c#14] parquet
scala> df.show()
+---+
| v|
+---+
+---+
// After this PR:
scala> df.explain(true)
== Analyzed Logical Plan ==
v: string
Project [v#43]
+- Filter a IN (list#42 [c#44])
: +- Project [v#45]
: +- Filter (c#46 = rpad(outer(c#44), 7, ))
: +- SubqueryAlias spark_catalog.default.t2
: +- Relation default.t2[v#45,c#46] parquet
+- SubqueryAlias spark_catalog.default.t1
+- Relation default.t1[v#43,c#44] parquet
scala> df.show()
+---+
| v|
+---+
| c|
+---+
```
This fix is lated to TPCDS q17; the query returns nothing because of this bug: https://github.com/apache/spark/pull/31886/files#r599333799
### Why are the changes needed?
Bugfix.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Unit tests added.
Closes#31940 from maropu/FixCharPadding.
Authored-by: Takeshi Yamamuro <yamamuro@apache.org>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
Currently, when implicit casting a data type to a `TypeCollection`, Spark returns the first convertible data type among `TypeCollection`.
In ANSI mode, we can make the behavior more reasonable by returning the closet convertible data type in `TypeCollection`.
In details, we first try to find the all the expected types we can implicitly cast:
1. if there is no convertible data types, return None;
2. if there is only one convertible data type, cast input as it;
3. otherwise if there are multiple convertible data types, find the closet data
type among them. If there is no such closet data type, return None.
Note that if the closet type is Float type and the convertible types contains Double type, simply return Double type as the closet type to avoid potential
precision loss on converting the Integral type as Float type.
### Why are the changes needed?
Make the type coercion rule for TypeCollection more reasonable and ANSI compatible.
E.g. returning Long instead of Double for`implicast(int, TypeCollect(Double, Long))`.
From ANSI SQL Spec section 4.33 "SQL-invoked routines"
Section 9.6 "Subject routine determination"
Section 10.4 "routine invocation"
### Does this PR introduce _any_ user-facing change?
Yes, in ANSI mode, implicit casting to a `TypeCollection` returns the narrowest convertible data type instead of the first convertible one.
### How was this patch tested?
Unit tests.
Closes#31859 from gengliangwang/implicitCastTypeCollection.
Lead-authored-by: Gengliang Wang <gengliang.wang@databricks.com>
Co-authored-by: Gengliang Wang <ltnwgl@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR fixes an issue that `col()`, `$"<name>"` and `df("name")` don't handle quoted column names like ``` `a``b.c` ```properly.
For example, if we have a following DataFrame.
```
val df1 = spark.sql("SELECT 'col1' AS `a``b.c`")
```
For the DataFrame, this query is successfully executed.
```
scala> df1.selectExpr("`a``b.c`").show
+-----+
|a`b.c|
+-----+
| col1|
+-----+
```
But the following query will fail because ``` df1("`a``b.c`") ``` throws an exception.
```
scala> df1.select(df1("`a``b.c`")).show
org.apache.spark.sql.AnalysisException: syntax error in attribute name: `a``b.c`;
at org.apache.spark.sql.catalyst.analysis.UnresolvedAttribute$.e$1(unresolved.scala:152)
at org.apache.spark.sql.catalyst.analysis.UnresolvedAttribute$.parseAttributeName(unresolved.scala:162)
at org.apache.spark.sql.catalyst.plans.logical.LogicalPlan.resolveQuoted(LogicalPlan.scala:121)
at org.apache.spark.sql.Dataset.resolve(Dataset.scala:221)
at org.apache.spark.sql.Dataset.col(Dataset.scala:1274)
at org.apache.spark.sql.Dataset.apply(Dataset.scala:1241)
... 49 elided
```
### Why are the changes needed?
It's a bug.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
New tests.
Closes#31854 from sarutak/fix-parseAttributeName.
Authored-by: Kousuke Saruta <sarutak@oss.nttdata.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
1. Add new expression `MultiplyYMInterval` which multiplies a `YearMonthIntervalType` expression by a `NumericType` expression including ByteType, ShortType, IntegerType, LongType, FloatType, DoubleType, DecimalType.
2. Extend binary arithmetic rules to support `numeric * year-month interval` and `year-month interval * numeric`.
### Why are the changes needed?
To conform the ANSI SQL standard which requires such operation over year-month intervals:
<img width="667" alt="Screenshot 2021-03-22 at 16 33 16" src="https://user-images.githubusercontent.com/1580697/111997810-77d1eb80-8b2c-11eb-951d-e43911d9c5db.png">
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
By running new tests:
```
$ build/sbt "test:testOnly *IntervalExpressionsSuite"
$ build/sbt "test:testOnly *ColumnExpressionSuite"
```
Closes#31929 from MaxGekk/interval-mul-div.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
forward-port https://github.com/apache/spark/pull/31811 to master
### What changes were proposed in this pull request?
For permanent views (and the new SQL temp view in Spark 3.1), we store the view SQL text and re-parse/analyze the view SQL text when reading the view. In the case of `SELECT * FROM ...`, we want to avoid view schema change (e.g. the referenced table changes its schema) and will record the view query output column names when creating the view, so that when reading the view we can add a `SELECT recorded_column_names FROM ...` to retain the original view query schema.
In Spark 3.1 and before, the final SELECT is added after the analysis phase: https://github.com/apache/spark/blob/branch-3.1/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/view.scala#L67
If the view query has duplicated output column names, we always pick the first column when reading a view. A simple repro:
```
scala> sql("create view c(x, y) as select 1 a, 2 a")
res0: org.apache.spark.sql.DataFrame = []
scala> sql("select * from c").show
+---+---+
| x| y|
+---+---+
| 1| 1|
+---+---+
```
In the master branch, we will fail at the view reading time due to b891862fb6 , which adds the final SELECT during analysis, so that the query fails with `Reference 'a' is ambiguous`
This PR proposes to resolve the view query output column names from the matching attributes by ordinal.
For example, `create view c(x, y) as select 1 a, 2 a`, the view query output column names are `[a, a]`. When we reading the view, there are 2 matching attributes (e.g.`[a#1, a#2]`) and we can simply match them by ordinal.
A negative example is
```
create table t(a int)
create view v as select *, 1 as col from t
replace table t(a int, col int)
```
When reading the view, the view query output column names are `[a, col]`, and there are two matching attributes of `col`, and we should fail the query. See the tests for details.
### Why are the changes needed?
bug fix
### Does this PR introduce _any_ user-facing change?
yes
### How was this patch tested?
new test
Closes#31930 from cloud-fan/view2.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This patch proposes to add a few public API change to DS v2, to make DS v2 scan can report metrics to Spark.
Two public interfaces are added.
* `CustomMetric`: metric interface at the driver side. It basically defines how Spark aggregates task metrics with the same metric name.
* `CustomTaskMetric`: task metric reported at executors. It includes a name and long value. Spark will collect these metric values and update internal metrics.
There are two public methods added to existing public interfaces. They are optional to DS v2 implementations.
* `PartitionReader.currentMetricsValues()`: returns an array of CustomTaskMetric. Here is where the actual metrics values are collected. Empty array by default.
* `Scan.supportedCustomMetrics()`: returns an array of supported custom metrics `CustomMetric`. Empty array by default.
### Why are the changes needed?
In order to report custom metrics, we need some public API change in DS v2 to make it possible.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
This only adds interfaces. In follow-up PRs where adding implementation there will be tests added. See #31451 and #31398 for some details and manual test there.
Closes#31476 from viirya/SPARK-34366.
Authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This patch proposes to disable fetching shuffle blocks in batch when io encryption is enabled. Adaptive Query Execution fetch contiguous shuffle blocks for the same map task in batch to reduce IO and improve performance. However, we found that batch fetching is incompatible with io encryption.
### Why are the changes needed?
Before this patch, we set `spark.io.encryption.enabled` to true, then run some queries which coalesced partitions by AEQ, may got following error message:
```14:05:52.638 WARN org.apache.spark.scheduler.TaskSetManager: Lost task 1.0 in stage 2.0 (TID 3) (11.240.37.88 executor driver): FetchFailed(BlockManagerId(driver, 11.240.37.88, 63574, None), shuffleId=0, mapIndex=0, mapId=0, reduceId=2, message=
org.apache.spark.shuffle.FetchFailedException: Stream is corrupted
at org.apache.spark.storage.ShuffleBlockFetcherIterator.throwFetchFailedException(ShuffleBlockFetcherIterator.scala:772)
at org.apache.spark.storage.BufferReleasingInputStream.read(ShuffleBlockFetcherIterator.scala:845)
at java.io.BufferedInputStream.fill(BufferedInputStream.java:246)
at java.io.BufferedInputStream.read(BufferedInputStream.java:265)
at java.io.DataInputStream.readInt(DataInputStream.java:387)
at org.apache.spark.sql.execution.UnsafeRowSerializerInstance$$anon$2$$anon$3.readSize(UnsafeRowSerializer.scala:113)
at org.apache.spark.sql.execution.UnsafeRowSerializerInstance$$anon$2$$anon$3.next(UnsafeRowSerializer.scala:129)
at org.apache.spark.sql.execution.UnsafeRowSerializerInstance$$anon$2$$anon$3.next(UnsafeRowSerializer.scala:110)
at scala.collection.Iterator$$anon$11.next(Iterator.scala:494)
at scala.collection.Iterator$$anon$10.next(Iterator.scala:459)
at org.apache.spark.util.CompletionIterator.next(CompletionIterator.scala:29)
at org.apache.spark.InterruptibleIterator.next(InterruptibleIterator.scala:40)
at scala.collection.Iterator$$anon$10.next(Iterator.scala:459)
at org.apache.spark.sql.execution.SparkPlan.$anonfun$getByteArrayRdd$1(SparkPlan.scala:345)
at org.apache.spark.rdd.RDD.$anonfun$mapPartitionsInternal$2(RDD.scala:898)
at org.apache.spark.rdd.RDD.$anonfun$mapPartitionsInternal$2$adapted(RDD.scala:898)
at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:52)
at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:373)
at org.apache.spark.rdd.RDD.iterator(RDD.scala:337)
at org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:90)
at org.apache.spark.scheduler.Task.run(Task.scala:131)
at org.apache.spark.executor.Executor$TaskRunner.$anonfun$run$3(Executor.scala:498)
at org.apache.spark.util.Utils$.tryWithSafeFinally(Utils.scala:1437)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:501)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
Caused by: java.io.IOException: Stream is corrupted
at net.jpountz.lz4.LZ4BlockInputStream.refill(LZ4BlockInputStream.java:200)
at net.jpountz.lz4.LZ4BlockInputStream.refill(LZ4BlockInputStream.java:226)
at net.jpountz.lz4.LZ4BlockInputStream.read(LZ4BlockInputStream.java:157)
at org.apache.spark.storage.BufferReleasingInputStream.read(ShuffleBlockFetcherIterator.scala:841)
... 25 more
)
```
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
New tests.
Closes#31898 from hezuojiao/fetch_shuffle_in_batch.
Authored-by: hezuojiao <hezuojiao@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Marked `RowNumberLike` and `RankLike` as not-nullable.
### Why are the changes needed?
`RowNumberLike` and `RankLike` SQL expressions never return null value. Marking them as non-nullable can have some performance benefits, because some optimizer rules apply only to non-nullable expressions
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Did not find any existing tests on the nullability of aggregate functions.
Plan stability suite partially covers this.
Closes#31924 from tanelk/SPARK-34812_nullability.
Authored-by: tanel.kiis@gmail.com <tanel.kiis@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Use fine-grained lock in SessionCatalog.tableExists, in order to lock currentDB variable rather than lock `tableExists` method which will block inner external catalog's behaviour.
### Why are the changes needed?
We have modified the underlying hive meta store which a different hive database is placed in its own shard for performance. However, we found that the synchronized lock limits the concurrency.
### How was this patch tested?
Existing tests.
Closes#31891 from woyumen4597/SPARK-34800.
Authored-by: woyumen4597 <woyumen4597@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Now that all the temporary views are wrapped with `TemporaryViewRelation`(#31273, #31652, and #31825), this PR proposes to update `SessionCatalog`'s APIs for temporary views to take or return more concrete types.
APIs that will take `TemporaryViewRelation` instead of `LogicalPlan`:
```
createTempView, createGlobalTempView, alterTempViewDefinition
```
APIs that will return `TemporaryViewRelation` instead of `LogicalPlan`:
```
getRawTempView, getRawGlobalTempView
```
APIs that will return `View` instead of `LogicalPlan`:
```
getTempView, getGlobalTempView, lookupTempView
```
### Why are the changes needed?
Internal refactoring to work with more concrete types.
### Does this PR introduce _any_ user-facing change?
No, this is internal refactoring.
### How was this patch tested?
Updated existing tests affected by the refactoring.
Closes#31906 from imback82/use_temporary_view_relation.
Authored-by: Terry Kim <yuminkim@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR fixes the `HybridRowQueue ` to respect the configured memory mode.
Besides, this PR also refactored the constructor of `MemoryConsumer` to accept the memory mode explicitly.
### Why are the changes needed?
`HybridRowQueue` supports both onHeap and offHeap manipulation. But it inherited the wrong `MemoryConsumer` constructor, which hard-coded the memory mode to `onHeap`.
### Does this PR introduce _any_ user-facing change?
No. (Maybe yes in some cases where users can't complete the job before could complete successfully after the fix because of `HybridRowQueue` is able to spill under offHeap mode now. )
### How was this patch tested?
Updated the existing test to make it test both offHeap and onHeap modes.
Closes#31152 from Ngone51/fix-MemoryConsumer-memorymode.
Authored-by: yi.wu <yi.wu@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
- Create a new rule `ResolveStreamWrite` for all analysis logic for streaming write.
- Add corresponding logical plans `WriteToStreamStatement` and `WriteToStream`.
### Why are the changes needed?
Currently, the analysis logic for streaming write is mixed in StreamingQueryManager. If we create a specific analyzer rule and separated logical plans, it should be helpful for further extension.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Existing tests.
Closes#31842 from xuanyuanking/SPARK-34748.
Authored-by: Yuanjian Li <yuanjian.li@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR fixes an incompatible behavior introduced by #31754.
The problem is that quoted name parts represented as a string are given to the constructor of `UnresolvedAttribute` which takes single string parameter, `sql` method invocation against the `UnresolvedAttrribute` returns different result than before.
One example is ``` UnresolvedAttribute("`a.b`").sql ```. This returned `a.b` before but it doesn't now.
See [this duscussion](https://github.com/apache/spark/pull/31754/files#r597181927) for more details.
### Why are the changes needed?
For compatibility.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
New assertion.
Closes#31885 from sarutak/followup-SPARK-34636.
Authored-by: Kousuke Saruta <sarutak@oss.nttdata.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Improve `PropagateEmptyRelation` to support join with false condition. For example:
```sql
SELECT * FROM t1 LEFT JOIN t2 ON false
```
Before this pr:
```
== Physical Plan ==
AdaptiveSparkPlan isFinalPlan=false
+- BroadcastNestedLoopJoin BuildRight, LeftOuter, false
:- FileScan parquet default.t1[a#4L]
+- BroadcastExchange IdentityBroadcastMode, [id=#40]
+- FileScan parquet default.t2[b#5L]
```
After this pr:
```
== Physical Plan ==
*(1) Project [a#4L, null AS b#5L]
+- *(1) ColumnarToRow
+- FileScan parquet default.t1[a#4L]
```
### Why are the changes needed?
Avoid `BroadcastNestedLoopJoin` to improve query performance.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Unit test.
Closes#31857 from wangyum/SPARK-28220.
Authored-by: Yuming Wang <yumwang@ebay.com>
Signed-off-by: Yuming Wang <yumwang@ebay.com>
### What changes were proposed in this pull request?
This patch proposes to fix a bug related to `NestedColumnAliasing`. The root cause is `Window` doesn't override `producedAttributes` so `NestedColumnAliasing` rule wrongly prune attributes produced by `Window`.
The master and branch-3.1 both have this issue.
### Why are the changes needed?
It is needed to fix a bug of nested column pruning.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Unit test.
Closes#31897 from viirya/SPARK-34776.
Authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
For all built-in datasources, prohibit saving of year-month and day-time intervals that were introduced by SPARK-27793. We plan to support saving of such types at the milestone 2, see SPARK-27790.
### Why are the changes needed?
To improve user experience with Spark SQL, and print nicer error message. Current error message might confuse users:
```
scala> Seq(java.time.Period.ofMonths(1)).toDF.write.mode("overwrite").json("/Users/maximgekk/tmp/123")
21/03/18 22:44:35 ERROR FileFormatWriter: Aborting job 8de402d7-ab69-4dc0-aa8e-14ef06bd2d6b.
org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in stage 1.0 failed 1 times, most recent failure: Lost task 0.0 in stage 1.0 (TID 1) (192.168.1.66 executor driver): org.apache.spark.SparkException: Task failed while writing rows.
at org.apache.spark.sql.errors.QueryExecutionErrors$.taskFailedWhileWritingRowsError(QueryExecutionErrors.scala:418)
at org.apache.spark.sql.execution.datasources.FileFormatWriter$.executeTask(FileFormatWriter.scala:298)
at org.apache.spark.sql.execution.datasources.FileFormatWriter$.$anonfun$write$15(FileFormatWriter.scala:211)
at org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:90)
at org.apache.spark.scheduler.Task.run(Task.scala:131)
at org.apache.spark.executor.Executor$TaskRunner.$anonfun$run$3(Executor.scala:498)
at org.apache.spark.util.Utils$.tryWithSafeFinally(Utils.scala:1437)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:501)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
Caused by: java.lang.RuntimeException: Failed to convert value 1 (class of class java.lang.Integer}) with the type of YearMonthIntervalType to JSON.
at scala.sys.package$.error(package.scala:30)
at org.apache.spark.sql.catalyst.json.JacksonGenerator.$anonfun$makeWriter$23(JacksonGenerator.scala:179)
at org.apache.spark.sql.catalyst.json.JacksonGenerator.$anonfun$makeWriter$23$adapted(JacksonGenerator.scala:176)
```
### Does this PR introduce _any_ user-facing change?
Yes. After the changes, the example above:
```
scala> Seq(java.time.Period.ofMonths(1)).toDF.write.mode("overwrite").json("/Users/maximgekk/tmp/123")
org.apache.spark.sql.AnalysisException: Cannot save interval data type into external storage.
```
### How was this patch tested?
1. Checked nested intervals:
```
scala> spark.range(1).selectExpr("""struct(timestamp'2021-01-02 00:01:02' - timestamp'2021-01-01 00:00:00')""").write.mode("overwrite").parquet("/Users/maximgekk/tmp/123")
org.apache.spark.sql.AnalysisException: Cannot save interval data type into external storage.
scala> Seq(Seq(java.time.Period.ofMonths(1))).toDF.write.mode("overwrite").json("/Users/maximgekk/tmp/123")
org.apache.spark.sql.AnalysisException: Cannot save interval data type into external storage.
```
2. By running existing test suites:
```
$ build/sbt -Phive-2.3 -Phive-thriftserver "test:testOnly *DataSourceV2DataFrameSuite"
$ build/sbt -Phive-2.3 -Phive-thriftserver "test:testOnly *DataSourceV2SQLSuite"
```
Closes#31884 from MaxGekk/ban-save-intervals.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
join condition 'a.attr == 'c.attr check the reference of these 2 objects which will always returns false. we need to use === instead
### Why are the changes needed?
Although this join condition always false doesn't break the test but it is not what we expected. We should fix it to avoid future confusing
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
UT
Closes#31890 from opensky142857/SPARK-34798.
Authored-by: Hongyi Zhang <hongyzhang@ebay.com>
Signed-off-by: Yuming Wang <yumwang@ebay.com>
### What changes were proposed in this pull request?
Change context classloader to Spark classloader at `RebaseDateTime.loadRebaseRecords`
### Why are the changes needed?
With custom `spark.sql.hive.metastore.version` and `spark.sql.hive.metastore.jars`.
Spark would use date formatter in `HiveShim` that convert `date` to `string`, if we set `spark.sql.legacy.timeParserPolicy=LEGACY` and the partition type is `date` the `RebaseDateTime` code will be invoked. At that moment, if `RebaseDateTime` is initialized the first time then context class loader is `IsolatedClientLoader`. Such error msg would throw:
```
java.lang.IllegalArgumentException: argument "src" is null
at com.fasterxml.jackson.databind.ObjectMapper._assertNotNull(ObjectMapper.java:4413)
at com.fasterxml.jackson.databind.ObjectMapper.readValue(ObjectMapper.java:3157)
at com.fasterxml.jackson.module.scala.ScalaObjectMapper.readValue(ScalaObjectMapper.scala:187)
at com.fasterxml.jackson.module.scala.ScalaObjectMapper.readValue$(ScalaObjectMapper.scala:186)
at org.apache.spark.sql.catalyst.util.RebaseDateTime$$anon$1.readValue(RebaseDateTime.scala:267)
at org.apache.spark.sql.catalyst.util.RebaseDateTime$.loadRebaseRecords(RebaseDateTime.scala:269)
at org.apache.spark.sql.catalyst.util.RebaseDateTime$.<init>(RebaseDateTime.scala:291)
at org.apache.spark.sql.catalyst.util.RebaseDateTime$.<clinit>(RebaseDateTime.scala)
at org.apache.spark.sql.catalyst.util.DateTimeUtils$.toJavaDate(DateTimeUtils.scala:109)
at org.apache.spark.sql.catalyst.util.LegacyDateFormatter.format(DateFormatter.scala:95)
at org.apache.spark.sql.catalyst.util.LegacyDateFormatter.format$(DateFormatter.scala:94)
at org.apache.spark.sql.catalyst.util.LegacySimpleDateFormatter.format(DateFormatter.scala:138)
at org.apache.spark.sql.hive.client.Shim_v0_13$ExtractableLiteral$1$.unapply(HiveShim.scala:661)
at org.apache.spark.sql.hive.client.Shim_v0_13.convert$1(HiveShim.scala:785)
at org.apache.spark.sql.hive.client.Shim_v0_13.$anonfun$convertFilters$4(HiveShim.scala:826)
```
```
java.lang.NoClassDefFoundError: Could not initialize class org.apache.spark.sql.catalyst.util.RebaseDateTime$
at org.apache.spark.sql.catalyst.util.DateTimeUtils$.toJavaDate(DateTimeUtils.scala:109)
at org.apache.spark.sql.catalyst.util.LegacyDateFormatter.format(DateFormatter.scala:95)
at org.apache.spark.sql.catalyst.util.LegacyDateFormatter.format$(DateFormatter.scala:94)
at org.apache.spark.sql.catalyst.util.LegacySimpleDateFormatter.format(DateFormatter.scala:138)
at org.apache.spark.sql.hive.client.Shim_v0_13$ExtractableLiteral$1$.unapply(HiveShim.scala:661)
at org.apache.spark.sql.hive.client.Shim_v0_13.convert$1(HiveShim.scala:785)
at org.apache.spark.sql.hive.client.Shim_v0_13.$anonfun$convertFilters$4(HiveShim.scala:826)
at scala.collection.immutable.Stream.flatMap(Stream.scala:493)
at org.apache.spark.sql.hive.client.Shim_v0_13.convertFilters(HiveShim.scala:826)
at org.apache.spark.sql.hive.client.Shim_v0_13.getPartitionsByFilter(HiveShim.scala:848)
at org.apache.spark.sql.hive.client.HiveClientImpl.$anonfun$getPartitionsByFilter$1(HiveClientImpl.scala:749)
at org.apache.spark.sql.hive.client.HiveClientImpl.$anonfun$withHiveState$1(HiveClientImpl.scala:291)
at org.apache.spark.sql.hive.client.HiveClientImpl.liftedTree1$1(HiveClientImpl.scala:224)
at org.apache.spark.sql.hive.client.HiveClientImpl.retryLocked(HiveClientImpl.scala:223)
at org.apache.spark.sql.hive.client.HiveClientImpl.withHiveState(HiveClientImpl.scala:273)
at org.apache.spark.sql.hive.client.HiveClientImpl.getPartitionsByFilter(HiveClientImpl.scala:747)
at org.apache.spark.sql.hive.HiveExternalCatalog.$anonfun$listPartitionsByFilter$1(HiveExternalCatalog.scala:1273)
```
The reproduce steps:
1. `spark.sql.hive.metastore.version` and `spark.sql.hive.metastore.jars`.
2. `CREATE TABLE t (c int) PARTITIONED BY (p date)`
3. `SET spark.sql.legacy.timeParserPolicy=LEGACY`
4. `SELECT * FROM t WHERE p='2021-01-01'`
### Does this PR introduce _any_ user-facing change?
Yes, bug fix.
### How was this patch tested?
pass `org.apache.spark.sql.catalyst.util.RebaseDateTimeSuite` and add new unit test to `HiveSparkSubmitSuite.scala`.
Closes#31864 from ulysses-you/SPARK-34772.
Authored-by: ulysses-you <ulyssesyou18@gmail.com>
Signed-off-by: Yuming Wang <yumwang@ebay.com>
### What changes were proposed in this pull request?
Support `timestamp +/- day-time interval`. In the PR, I propose to extend the `TimeAdd` expression and support `DayTimeIntervalType` as the `interval` parameter. The expression invokes the new method `DateTimeUtils.timestampAddDayTime()` which splits the input day-time interval to `days` and `microsecond adjustment` of a day, and adds `days` (and the microseconds) to a local timestamp derived from the given timestamp at the given time zone. The resulted local timestamp is converted back to the offset in microseconds since the epoch.
Also I updated the rules that handle `CalendarIntervalType` and produce `TimeAdd` to take into account new type `DateTimeIntervalType` for the `interval` parameter of `TimeAdd`.
### Why are the changes needed?
To conform the ANSI SQL standard which requires to support such operation over timestamps and intervals:
<img width="811" alt="Screenshot 2021-03-12 at 11 36 14" src="https://user-images.githubusercontent.com/1580697/111081674-865d4900-8515-11eb-86c8-3538ecaf4804.png">
### Does this PR introduce _any_ user-facing change?
Should not since new intervals have not been released yet.
### How was this patch tested?
By running new tests:
```
$ build/sbt "test:testOnly *DateTimeUtilsSuite"
$ build/sbt "test:testOnly *DateExpressionsSuite"
$ build/sbt "test:testOnly *ColumnExpressionSuite"
```
Closes#31855 from MaxGekk/timestamp-add-day-time-interval.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR proposes to deduplicate the source table when there're conflicting attributes between the target table and the source table.
### Why are the changes needed?
When resolving the `UpdateAction`, which could reference attributes from both target and source tables, Spark should know clearly where the attribute comes from when there're conflicting attributes instead of picking up a random one.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Added a unit test and updated existing tests.
Closes#31835 from Ngone51/dedup-MergeIntoTable.
Authored-by: yi.wu <yi.wu@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Remove all SQLConf.get to conf if extends from SQLConfHelper
### Why are the changes needed?
Clean up code.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existing unit tests.
Closes#31822 from leoluan2009/SPARK-34728.
Authored-by: Luan <luanxuedong2009@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR group exception messages in `/core/src/main/scala/org/apache/spark/sql/execution/datasources`.
### Why are the changes needed?
It will largely help with standardization of error messages and its maintenance.
### Does this PR introduce _any_ user-facing change?
No. Error messages remain unchanged.
### How was this patch tested?
No new tests - pass all original tests to make sure it doesn't break any existing behavior.
Closes#31757 from beliefer/SPARK-33602.
Authored-by: gengjiaan <gengjiaan@360.cn>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR simplifies `Analyzer.resolveLiteralFunction` to always create the `Alias`. The caller side will remove the `Alias` if it's not necessary.
### Why are the changes needed?
code simplification.
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
existing tests
Closes#31844 from cloud-fan/minor.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
This is a follow-up of https://github.com/apache/spark/pull/31808 and simplifies its fix to one line (excluding comments).
### Why are the changes needed?
code simplification
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
N/A
Closes#31843 from cloud-fan/simplify.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Takeshi Yamamuro <yamamuro@apache.org>
### What changes were proposed in this pull request?
This PR proposes to follow Univocity's input buffer.
### Why are the changes needed?
- Firstly, it's best to trust their judgement on the default values. Also 128 is too low.
- Default values arguably have more test coverage in Univocity.
- It will also fix https://github.com/uniVocity/univocity-parsers/issues/449
- ^ is a regression compared to Spark 2.4
### Does this PR introduce _any_ user-facing change?
No. In addition, It fixes a regression.
### How was this patch tested?
Manually tested, and added a unit test.
Closes#31858 from HyukjinKwon/SPARK-34768.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR updates `InMemoryCatalog.tableExists` to return false if database doesn't exist, instead of failing. The new behavior is consistent with `HiveExternalCatalog` which is used in production, so this bug mostly only affects tests.
### Why are the changes needed?
bug fix
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
a new test
Closes#31860 from cloud-fan/catalog.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Push down limit through `Window` when the partitionSpec of all window functions is empty and the same order is used. This is a real case from production:
This pr support 2 cases:
1. All window functions have same orderSpec:
```sql
SELECT *, ROW_NUMBER() OVER(ORDER BY a) AS rn, RANK() OVER(ORDER BY a) AS rk FROM t1 LIMIT 5;
== Optimized Logical Plan ==
Window [row_number() windowspecdefinition(a#9L ASC NULLS FIRST, specifiedwindowframe(RowFrame, unboundedpreceding$(), currentrow$())) AS rn#4, rank(a#9L) windowspecdefinition(a#9L ASC NULLS FIRST, specifiedwindowframe(RowFrame, unboundedpreceding$(), currentrow$())) AS rk#5], [a#9L ASC NULLS FIRST]
+- GlobalLimit 5
+- LocalLimit 5
+- Sort [a#9L ASC NULLS FIRST], true
+- Relation default.t1[A#9L,B#10L,C#11L] parquet
```
2. There is a window function with a different orderSpec:
```sql
SELECT a, ROW_NUMBER() OVER(ORDER BY a) AS rn, RANK() OVER(ORDER BY b DESC) AS rk FROM t1 LIMIT 5;
== Optimized Logical Plan ==
Project [a#9L, rn#4, rk#5]
+- Window [rank(b#10L) windowspecdefinition(b#10L DESC NULLS LAST, specifiedwindowframe(RowFrame, unboundedpreceding$(), currentrow$())) AS rk#5], [b#10L DESC NULLS LAST]
+- GlobalLimit 5
+- LocalLimit 5
+- Sort [b#10L DESC NULLS LAST], true
+- Window [row_number() windowspecdefinition(a#9L ASC NULLS FIRST, specifiedwindowframe(RowFrame, unboundedpreceding$(), currentrow$())) AS rn#4], [a#9L ASC NULLS FIRST]
+- Project [a#9L, b#10L]
+- Relation default.t1[A#9L,B#10L,C#11L] parquet
```
### Why are the changes needed?
Improve query performance.
```scala
spark.range(500000000L).selectExpr("id AS a", "id AS b").write.saveAsTable("t1")
spark.sql("SELECT *, ROW_NUMBER() OVER(ORDER BY a) AS rowId FROM t1 LIMIT 5").show
```
Before this pr | After this pr
-- | --
 | 
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Unit test.
Closes#31691 from wangyum/SPARK-34575.
Authored-by: Yuming Wang <yumwang@ebay.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
For the following cases, ABS should throw exceptions since the results are out of the range of the result data types in ANSI mode.
```
SELECT abs(${Int.MinValue});
SELECT abs(${Long.MinValue});
```
### Why are the changes needed?
Better ANSI compliance
### Does this PR introduce _any_ user-facing change?
Yes, Abs throws an exception if input is out of range in ANSI mode
### How was this patch tested?
Unit test
Closes#31836 from gengliangwang/ansiAbs.
Authored-by: Gengliang Wang <gengliang.wang@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
For DDL commands like DROP VIEW, they don't really need to resolve the view (parse and analyze the view SQL text), they just need to get the view metadata.
This PR fixes the rule `ResolveTempViews` to only resolve the temp view for `UnresolvedRelation`. This also fixes a bug for DROP VIEW, as previously it tried to resolve the view and failed to drop invalid views.
### Why are the changes needed?
bug fix
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
new test
Closes#31853 from cloud-fan/view-resolve.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR removes one unused variable in `NewInstance.constructor`.
### Why are the changes needed?
This looks like a variable for debugging at the initial commit of SPARK-23584 .
- 1b08c4393c (diff-2a36e31684505fd22e2d12a864ce89fd350656d716a3f2d7789d2cdbe38e15fbR461)
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Pass the CIs.
Closes#31838 from dongjoon-hyun/minor-object.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Support `timestamp +/- year-month interval`. In the PR, I propose to introduce new binary expression `TimestampAddYMInterval` similarly to `DateAddYMInterval`. It invokes new method `timestampAddMonths` from `DateTimeUtils` by passing a timestamp as an offset in microseconds since the epoch, amount of months from the giveb year-month interval, and the time zone ID in which the operation is performed. The `timestampAddMonths()` method converts the input microseconds to a local timestamp, adds months to it, and converts the results back to an instant in microseconds at the given time zone.
### Why are the changes needed?
To conform the ANSI SQL standard which requires to support such operation over timestamps and intervals:
<img width="811" alt="Screenshot 2021-03-12 at 11 36 14" src="https://user-images.githubusercontent.com/1580697/111081674-865d4900-8515-11eb-86c8-3538ecaf4804.png">
### Does this PR introduce _any_ user-facing change?
Should not since new intervals have not been released yet.
### How was this patch tested?
By running new tests:
```
$ build/sbt "test:testOnly *DateTimeUtilsSuite"
$ build/sbt "test:testOnly *DateExpressionsSuite"
$ build/sbt "test:testOnly *ColumnExpressionSuite"
```
Closes#31832 from MaxGekk/timestamp-add-year-month-interval.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
This PR aims to make `ExpressionEncoderSuite` to use `deepEquals` instead of `equals` when `input` is `array of array`.
This comparison code itself was added by SPARK-11727 at Apache Spark 1.6.0.
### Why are the changes needed?
Currently, the interpreted mode fails for `array of array` because the following line is used.
```
Arrays.equals(b1.asInstanceOf[Array[AnyRef]], b2.asInstanceOf[Array[AnyRef]])
```
### Does this PR introduce _any_ user-facing change?
No. This is a test-only PR.
### How was this patch tested?
Pass the existing CIs.
Closes#31837 from dongjoon-hyun/SPARK-34743.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
In `Analyzer.resolveExpression`, we have a parameter to decide if we should remove unnecessary `Alias` or not. This is over complicated and we can always remove unnecessary `Alias`.
This PR simplifies this part and removes the parameter.
### Why are the changes needed?
code cleanup
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
existing tests
Closes#31758 from cloud-fan/resolve.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
In the PR, I propose to cast the input float to double in the `SecondsToTimestamp` expression in the same way as in the `Cast` expression.
### Why are the changes needed?
To have the same results from `CAST(<float> AS TIMESTAMP)` and from `TIMESTAMP_SECONDS`:
```sql
spark-sql> SELECT CAST(16777215.0f AS TIMESTAMP);
1970-07-14 07:20:15
spark-sql> SELECT TIMESTAMP_SECONDS(16777215.0f);
1970-07-14 07:20:14.951424
```
### Does this PR introduce _any_ user-facing change?
Yes. After the changes:
```sql
spark-sql> SELECT TIMESTAMP_SECONDS(16777215.0f);
1970-07-14 07:20:15
```
### How was this patch tested?
By running new test:
```
$ build/sbt "test:testOnly *DateExpressionsSuite"
```
Closes#31831 from MaxGekk/adjust-SecondsToTimestamp.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
In non-ANSI mode, casting float to timestamp has different implementation for codegen on and off.
Codegen on:
1. Multiply float input by MICROS_PER_SECOND
2. Cast resulting float value to long
Codegen off:
1. CAST float input to double input
2. Multiply double input by MICROS_PER_SECOND
3. Cast resulting double value to long
In the PR, I propose to align to non-codegen code, and cast input float to double in codegen.
### Why are the changes needed?
This fixes the issue which is demonstrated by the code:
```sql
spark-sql> CREATE TEMP VIEW v1 AS SELECT 16777215.0f AS f;
spark-sql> SELECT * FROM v1;
1.6777215E7
spark-sql> SELECT CAST(f AS TIMESTAMP) FROM v1;
1970-07-14 07:20:15
spark-sql> CACHE TABLE v1;
spark-sql> SELECT * FROM v1;
1.6777215E7
spark-sql> SELECT CAST(f AS TIMESTAMP) FROM v1;
1970-07-14 07:20:14.951424
```
The result from the cached view **1970-07-14 07:20:14.951424** is different from un-cached view **1970-07-14 07:20:15**.
### Does this PR introduce _any_ user-facing change?
Yes. After the changes, the example above outputs the same timestamp for the cached view:
```sql
spark-sql> CACHE TABLE v1;
spark-sql> SELECT * FROM v1;
1.6777215E7
spark-sql> SELECT CAST(f AS TIMESTAMP) FROM v1;
1970-07-14 07:20:15
```
### How was this patch tested?
By running new test:
```
$ build/sbt "test:testOnly *CastSuite"
```
Closes#31819 from MaxGekk/fix-float-to-timestamp.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This patch proposes to fix incorrect parameter type for subexpression elimination under whole-stage.
### Why are the changes needed?
If the parameter is a byte array, the subexpression elimination under wholestage codegen will use incorrect parameter type and cause compile error. Although Spark can automatically fallback to interpreted mode, we should fix it.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Manually test with customer application. Unit test.
Closes#31814 from viirya/SPARK-34723.
Authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Signed-off-by: Liang-Chi Hsieh <viirya@gmail.com>
### What changes were proposed in this pull request?
Support `date +/- year-month interval`. In the PR, I propose to re-use existing code from the `AddMonths` expression, and extract it to the common base class `AddMonthsBase`. That base class is used in new expression `DateAddYMInterval` and in the existing one `AddMonths` (the `add_months` function).
### Why are the changes needed?
To conform the ANSI SQL standard which requires to support such operation over dates and intervals:
<img width="811" alt="Screenshot 2021-03-12 at 11 36 14" src="https://user-images.githubusercontent.com/1580697/110914390-5f412480-8327-11eb-9f8b-e92e73c0b9cd.png">
### Does this PR introduce _any_ user-facing change?
Should not since new intervals have not been released yet.
### How was this patch tested?
By running new tests:
```
$ build/sbt "test:testOnly *ColumnExpressionSuite"
$ build/sbt "test:testOnly *DateExpressionsSuite"
```
Closes#31812 from MaxGekk/date-add-year-month-interval.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This bug was introduced by SPARK-23583 at Apache Spark 2.4.0.
This PR aims to use `getMethod` instead of `getDeclaredMethod`.
```scala
- obj.getClass.getDeclaredMethod(functionName, argClasses: _*)
+ obj.getClass.getMethod(functionName, argClasses: _*)
```
### Why are the changes needed?
`getDeclaredMethod` does not search the super class's method. To invoke `GenericArrayData.toIntArray`, we need to use `getMethod` because it's declared at the super class `ArrayData`.
```
[info] - encode/decode for array of int: [I74655d03 (interpreted path) *** FAILED *** (14 milliseconds)
[info] Exception thrown while decoding
[info] Converted: [0,1000000020,3,0,ffffff850000001f,4]
[info] Schema: value#680
[info] root
[info] -- value: array (nullable = true)
[info] |-- element: integer (containsNull = false)
[info]
[info]
[info] Encoder:
[info] class[value[0]: array<int>] (ExpressionEncoderSuite.scala:578)
[info] org.scalatest.exceptions.TestFailedException:
[info] at org.scalatest.Assertions.newAssertionFailedException(Assertions.scala:472)
[info] at org.scalatest.Assertions.newAssertionFailedException$(Assertions.scala:471)
[info] at org.scalatest.funsuite.AnyFunSuite.newAssertionFailedException(AnyFunSuite.scala:1563)
[info] at org.scalatest.Assertions.fail(Assertions.scala:949)
[info] at org.scalatest.Assertions.fail$(Assertions.scala:945)
[info] at org.scalatest.funsuite.AnyFunSuite.fail(AnyFunSuite.scala:1563)
[info] at org.apache.spark.sql.catalyst.encoders.ExpressionEncoderSuite.$anonfun$encodeDecodeTest$1(ExpressionEncoderSuite.scala:578)
[info] at org.apache.spark.sql.catalyst.encoders.ExpressionEncoderSuite.verifyNotLeakingReflectionObjects(ExpressionEncoderSuite.scala:656)
[info] at org.apache.spark.sql.catalyst.encoders.ExpressionEncoderSuite.$anonfun$testAndVerifyNotLeakingReflectionObjects$2(ExpressionEncoderSuite.scala:669)
[info] at org.apache.spark.sql.catalyst.plans.CodegenInterpretedPlanTest.$anonfun$test$4(PlanTest.scala:50)
[info] at org.apache.spark.sql.catalyst.plans.SQLHelper.withSQLConf(SQLHelper.scala:54)
[info] at org.apache.spark.sql.catalyst.plans.SQLHelper.withSQLConf$(SQLHelper.scala:38)
[info] at org.apache.spark.sql.catalyst.encoders.ExpressionEncoderSuite.withSQLConf(ExpressionEncoderSuite.scala:118)
[info] at org.apache.spark.sql.catalyst.plans.CodegenInterpretedPlanTest.$anonfun$test$3(PlanTest.scala:50)
...
[info] Cause: java.lang.RuntimeException: Error while decoding: java.lang.NoSuchMethodException: org.apache.spark.sql.catalyst.util.GenericArrayData.toIntArray()
[info] mapobjects(lambdavariable(MapObject, IntegerType, false, -1), assertnotnull(lambdavariable(MapObject, IntegerType, false, -1)), input[0, array<int>, true], None).toIntArray
[info] at org.apache.spark.sql.catalyst.encoders.ExpressionEncoder$Deserializer.apply(ExpressionEncoder.scala:186)
[info] at org.apache.spark.sql.catalyst.encoders.ExpressionEncoderSuite.$anonfun$encodeDecodeTest$1(ExpressionEncoderSuite.scala:576)
[info] at org.apache.spark.sql.catalyst.encoders.ExpressionEncoderSuite.verifyNotLeakingReflectionObjects(ExpressionEncoderSuite.scala:656)
[info] at org.apache.spark.sql.catalyst.encoders.ExpressionEncoderSuite.$anonfun$testAndVerifyNotLeakingReflectionObjects$2(ExpressionEncoderSuite.scala:669)
```
### Does this PR introduce _any_ user-facing change?
This causes a runtime exception when we use the interpreted mode.
### How was this patch tested?
Pass the modified unit test case.
Closes#31816 from dongjoon-hyun/SPARK-34724.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR fixes an issue that `sql` method in the following classes which take qualified names don't quote the qualified names properly.
* UnresolvedAttribute
* AttributeReference
* Alias
One instance caused by this issue is reported in SPARK-34626.
```
UnresolvedAttribute("a" :: "b" :: Nil).sql
`a.b` // expected: `a`.`b`
```
And other instances are like as follows.
```
UnresolvedAttribute("a`b"::"c.d"::Nil).sql
a`b.`c.d` // expected: `a``b`.`c.d`
AttributeReference("a.b", IntegerType)(qualifier = "c.d"::Nil).sql
c.d.`a.b` // expected: `c.d`.`a.b`
Alias(AttributeReference("a", IntegerType)(), "b.c")(qualifier = "d.e"::Nil).sql
`a` AS d.e.`b.c` // expected: `a` AS `d.e`.`b.c`
```
### Why are the changes needed?
This is a bug.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
New test.
Closes#31754 from sarutak/fix-qualified-names.
Authored-by: Kousuke Saruta <sarutak@oss.nttdata.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
In the PR, I propose to override the `typeName()` method in `YearMonthIntervalType` and `DayTimeIntervalType`, and assign them names according to the ANSI SQL standard:
<img width="836" alt="Screenshot 2021-03-11 at 17 29 04" src="https://user-images.githubusercontent.com/1580697/110802854-a54aa980-828f-11eb-956d-dd4fbf14aa72.png">
but keep the type name as singular according existing naming convention for other types.
### Why are the changes needed?
To improve Spark SQL user experience, and have readable types in error messages.
### Does this PR introduce _any_ user-facing change?
Should not since the types has not been released yet.
### How was this patch tested?
By running the modified tests:
```
$ build/sbt "test:testOnly *ExpressionTypeCheckingSuite"
$ build/sbt "sql/testOnly *SQLQueryTestSuite -- -z windowFrameCoercion.sql"
$ build/sbt "sql/testOnly *SQLQueryTestSuite -- -z literals.sql"
```
Closes#31810 from MaxGekk/interval-types-name.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This is a bug caused by https://issues.apache.org/jira/browse/SPARK-31670 . We remove the `Alias` when resolving column references in grouping expressions, which breaks `ResolveCreateNamedStruct`
### Why are the changes needed?
bug fix
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
new tests
Closes#31808 from cloud-fan/bug.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
In the PR, I propose to especially handle the amount of seconds `-9223372036855` in `IntervalUtils. durationToMicros()`. Starting from the amount (any durations with the second field < `-9223372036855`), input durations cannot fit to `Long` in the conversion to microseconds. For example, the amount of microseconds = `Long.MinValue = -9223372036854775808` can be represented in two forms:
1. seconds = -9223372036854, nanoAdjustment = -775808, or
2. seconds = -9223372036855, nanoAdjustment = +224192
And the method `Duration.ofSeconds()` produces the last form but such form causes overflow while converting `-9223372036855` seconds to microseconds.
In the PR, I propose to convert the second form to the first one if the second field of input duration is equal to `-9223372036855`.
### Why are the changes needed?
The changes fix the issue demonstrated by the code:
```scala
scala> durationToMicros(microsToDuration(Long.MinValue))
java.lang.ArithmeticException: long overflow
at java.lang.Math.multiplyExact(Math.java:892)
at org.apache.spark.sql.catalyst.util.IntervalUtils$.durationToMicros(IntervalUtils.scala:782)
... 49 elided
```
The `durationToMicros()` method cannot handle valid output of `microsToDuration()`.
### Does this PR introduce _any_ user-facing change?
Should not since new interval types has not been released yet.
### How was this patch tested?
By running new UT from `IntervalUtilsSuite`.
Closes#31799 from MaxGekk/fix-min-duration.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Add `Not(In)` and `Not(InSet)` pattern when convert filter to metastore.
### Why are the changes needed?
`NOT IN` is a useful condition to prune partition, it would be better to support it.
Technically, we can convert `c not in(x,y)` to `c != x and c != y`, then push it to metastore.
Avoid metastore overflow and respect the config `spark.sql.hive.metastorePartitionPruningInSetThreshold`, `Not(InSet)` won't push to metastore if it's value exceeds the threshold.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Add test.
Closes#31646 from ulysses-you/SPARK-34538.
Authored-by: ulysses-you <ulyssesyou18@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Extend the `Add`, `Subtract` and `UnaryMinus` expression to support `DayTimeIntervalType` and `YearMonthIntervalType` added by #31614.
Note: the expressions can throw the `overflow` exception independently from the SQL config `spark.sql.ansi.enabled`. In this way, the modified expressions always behave in the ANSI mode for the intervals.
### Why are the changes needed?
To conform to the ANSI SQL standard which defines `-/+` over intervals:
<img width="822" alt="Screenshot 2021-03-09 at 21 59 22" src="https://user-images.githubusercontent.com/1580697/110523128-bd50ea80-8122-11eb-9982-782da0088d27.png">
### Does this PR introduce _any_ user-facing change?
Should not since new types have not been released yet.
### How was this patch tested?
By running new tests in the test suites:
```
$ build/sbt "test:testOnly *ArithmeticExpressionSuite"
$ build/sbt "test:testOnly *ColumnExpressionSuite"
```
Closes#31789 from MaxGekk/add-subtruct-intervals.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
SPARK-23596 added `CodegenInterpretedPlanTest` at Apache Spark 2.4.0 in a wrong way because `withSQLConf` depends on the execution time `SQLConf.get` instead of `test` function declaration time. So, the following code executes the test twice without controlling the `CodegenObjectFactoryMode`. This PR aims to fix it correct and introduce a new function `testFallback`.
```scala
trait CodegenInterpretedPlanTest extends PlanTest {
override protected def test(
testName: String,
testTags: Tag*)(testFun: => Any)(implicit pos: source.Position): Unit = {
val codegenMode = CodegenObjectFactoryMode.CODEGEN_ONLY.toString
val interpretedMode = CodegenObjectFactoryMode.NO_CODEGEN.toString
withSQLConf(SQLConf.CODEGEN_FACTORY_MODE.key -> codegenMode) {
super.test(testName + " (codegen path)", testTags: _*)(testFun)(pos)
}
withSQLConf(SQLConf.CODEGEN_FACTORY_MODE.key -> interpretedMode) {
super.test(testName + " (interpreted path)", testTags: _*)(testFun)(pos)
}
}
}
```
### Why are the changes needed?
1. We need to use like the following.
```scala
super.test(testName + " (codegen path)", testTags: _*)(
withSQLConf(SQLConf.CODEGEN_FACTORY_MODE.key -> codegenMode) { testFun })(pos)
super.test(testName + " (interpreted path)", testTags: _*)(
withSQLConf(SQLConf.CODEGEN_FACTORY_MODE.key -> interpretedMode) { testFun })(pos)
```
2. After we fix this behavior with the above code, several test cases including SPARK-34596 and SPARK-34607 fail because they didn't work at both `CODEGEN` and `INTERPRETED` mode. Those test cases only work at `FALLBACK` mode. So, inevitably, we need to introduce `testFallback`.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Pass the CIs.
Closes#31766 from dongjoon-hyun/SPARK-34596-SPARK-34607.
Lead-authored-by: Dongjoon Hyun <dhyun@apple.com>
Co-authored-by: Dongjoon Hyun <dongjoon@apache.org>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This PR adds a default null ordering to public `SortDirection` to match the Catalyst behavior.
### Why are the changes needed?
The SQL standard does not define the default null ordering for a sort direction. That's why it is up to a query engine to assign one. We need to standardize this in our public connector expressions to avoid ambiguity. That's why I propose to match the behavior in our Catalyst expressions.
### Does this PR introduce _any_ user-facing change?
Yes, it affects unreleased connector expression API.
### How was this patch tested?
Existing tests.
Closes#31580 from aokolnychyi/spark-34457.
Authored-by: Anton Okolnychyi <aokolnychyi@apple.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR proposes the following:
* `AlterViewAs.query` is currently analyzed in the physical operator `AlterViewAsCommand`, but it should be analyzed during the analysis phase.
* When `spark.sql.legacy.storeAnalyzedPlanForView` is set to true, store `TermporaryViewRelation` which wraps the analyzed plan, similar to #31273.
* Try to uncache the view you are altering.
### Why are the changes needed?
Analyzing a plan should be done in the analysis phase if possible.
Not uncaching the view (existing behavior) seems like a bug since the cache may not be used again.
### Does this PR introduce _any_ user-facing change?
Yes, now the view can be uncached if it's already cached.
### How was this patch tested?
Added new tests around uncaching.
The existing tests such as `SQLViewSuite` should cover the analysis changes.
Closes#31652 from imback82/alter_view_child.
Authored-by: Terry Kim <yuminkim@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR intends to remove unnecessary `SQLConf.withExistingConf` in `CastSuite`; since we've remove `ParVector ` in #31775, we no longer need to copy SQL configs into each thread env.
### Why are the changes needed?
Clean up the code.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Run the existing tests.
Closes#31785 from maropu/UpdateCastSuite.
Authored-by: Takeshi Yamamuro <yamamuro@apache.org>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Add `DayTimeIntervalType` and `YearMonthIntervalType` to `DataTypeTestUtils.ordered`/`atomicTypes`, and implement values generation of those types in `LiteralGenerator`/`RandomDataGenerator`. In this way, the types will be tested automatically in:
1. ArithmeticExpressionSuite:
- "function least"
- "function greatest"
2. PredicateSuite
- "BinaryComparison consistency check"
- "AND, OR, EqualTo, EqualNullSafe consistency check"
3. ConditionalExpressionSuite
- "if"
4. RandomDataGeneratorSuite
- "Basic types"
5. CastSuite
- "null cast"
- "up-cast"
- "SPARK-27671: cast from nested null type in struct"
6. OrderingSuite
- "GenerateOrdering with DayTimeIntervalType"
- "GenerateOrdering with YearMonthIntervalType"
7. PredicateSuite
- "IN with different types"
8. UnsafeRowSuite
- "calling get(ordinal, datatype) on null columns"
9. SortSuite
- "sorting on YearMonthIntervalType ..."
- "sorting on DayTimeIntervalType ..."
### Why are the changes needed?
To improve test coverage.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
By running the affected test suites.
Closes#31782 from MaxGekk/test-interval-as-atomic.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR updates UnresolvedTableValuedFunction's name to be a FunctionIdentifier instead of a string.
### Why are the changes needed?
To make UnresolvedTableValuedFunction consistent with UnresolvedFunction that uses FunctionIdentifier as the function name.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Unit test.
Closes#31749 from allisonwang-db/spark-34627.
Authored-by: allisonwang-db <66282705+allisonwang-db@users.noreply.github.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
RewritePredicateSubquery Optimizer Rule must not update Filters without subqueries.
Following is one such example.
```
=== Applying Rule org.apache.spark.sql.catalyst.optimizer.RewritePredicateSubquery ===
Project [a#0] Project [a#0]
!+- Filter (((a#0 > 1) OR (b#1 > 2)) AND ((c#2 > 1) AND (d#3 > 2))) +- Filter ((((a#0 > 1) OR (b#1 > 2)) AND (c#2 > 1)) AND (d#3 > 2))
+- LocalRelation <empty>, [a#0, b#1, c#2, d#3] +- LocalRelation <empty>, [a#0, b#1, c#2, d#3]
```
### Why are the changes needed?
minor change.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existing UTs pass.
Closes#31712 from Swinky/rewritePredicateFix.
Authored-by: Swinky <mannswinky@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
In the PR, I propose to extend Spark SQL API to accept [`java.time.Period`](https://docs.oracle.com/javase/8/docs/api/java/time/Period.html) as an external type of recently added new Catalyst type - `YearMonthIntervalType` (see #31614). The Java class `java.time.Period` has similar semantic to ANSI SQL year-month interval type, and it is the most suitable to be an external type for `YearMonthIntervalType`. In more details:
1. Added `PeriodConverter` which converts `java.time.Period` instances to/from internal representation of the Catalyst type `YearMonthIntervalType` (to `Int` type). The `PeriodConverter` object uses new methods of `IntervalUtils`:
- `periodToMonths()` converts the input period to the total length in months. If this period is too large to fit `Int`, the method throws the exception `ArithmeticException`. **Note:** _the input period has "days" precision, the method just ignores the days unit._
- `monthToPeriod()` obtains a `java.time.Period` representing a number of months.
2. Support new type `YearMonthIntervalType` in `RowEncoder` via the methods `createDeserializerForPeriod()` and `createSerializerForJavaPeriod()`.
3. Extended the Literal API to construct literals from `java.time.Period` instances.
### Why are the changes needed?
1. To allow users parallelization of `java.time.Period` collections, and construct year-month interval columns. Also to collect such columns back to the driver side.
2. This will allow to write tests in other sub-tasks of SPARK-27790.
### Does this PR introduce _any_ user-facing change?
The PR extends existing functionality. So, users can parallelize instances of the `java.time.Duration` class and collect them back:
```scala
scala> val ds = Seq(java.time.Period.ofYears(10).withMonths(2)).toDS
ds: org.apache.spark.sql.Dataset[java.time.Period] = [value: yearmonthinterval]
scala> ds.collect
res0: Array[java.time.Period] = Array(P10Y2M)
```
### How was this patch tested?
- Added a few tests to `CatalystTypeConvertersSuite` to check conversion from/to `java.time.Period`.
- Checking row encoding by new tests in `RowEncoderSuite`.
- Making literals of `YearMonthIntervalType` are tested in `LiteralExpressionSuite`.
- Check collecting by `DatasetSuite` and `JavaDatasetSuite`.
- New tests in `IntervalUtilsSuites` to check conversions `java.time.Period` <-> months.
Closes#31765 from MaxGekk/java-time-period.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
