### What changes were proposed in this pull request?
1. Add the SQL config `spark.sql.legacy.parquet.int96RebaseModeInWrite` to control timestamps rebasing in saving them as INT96. It supports the same set of values as `spark.sql.legacy.parquet.datetimeRebaseModeInWrite` but the default value is `LEGACY` to preserve backward compatibility with Spark <= 3.0.
2. Write the metadata key `org.apache.spark.int96NoRebase` to parquet files if the files are saved with `spark.sql.legacy.parquet.int96RebaseModeInWrite` isn't set to `LEGACY`.
3. Add the SQL config `spark.sql.legacy.parquet.datetimeRebaseModeInRead` to control loading INT96 timestamps when parquet metadata doesn't have enough info (the `org.apache.spark.int96NoRebase` tag) about parquet writer - either INT96 was written by Proleptic Gregorian system or some Julian one.
4. Modified Vectorized and Parquet-mr Readers to support loading/saving INT96 timestamps w/o rebasing depending on SQL config and the metadata tag:
- **No rebasing** in testing when the SQL config `spark.test.forceNoRebase` is set to `true`
- **No rebasing** if parquet metadata contains the tag `org.apache.spark.int96NoRebase`. This is the case when parquet files are saved by Spark >= 3.1 with `spark.sql.legacy.parquet.datetimeRebaseModeInWrite` is set to `CORRECTED`, or saved by other systems with the tag `org.apache.spark.int96NoRebase`.
- **With rebasing** if parquet files saved by Spark (any versions) without the metadata tag `org.apache.spark.int96NoRebase`.
- Rebasing depend on the SQL config `spark.sql.legacy.parquet.datetimeRebaseModeInRead` if there are no metadata tags `org.apache.spark.version` and `org.apache.spark.int96NoRebase`.
New SQL configs are added instead of re-using existing `spark.sql.legacy.parquet.datetimeRebaseModeInWrite` and `spark.sql.legacy.parquet.datetimeRebaseModeInRead` because of:
- To allow users have different modes for INT96 and for TIMESTAMP_MICROS (MILLIS). For example, users might want to save INT96 as LEGACY but TIMESTAMP_MICROS as CORRECTED.
- To have different modes for INT96 and DATE in load (or in save).
- To be backward compatible with Spark 2.4. For now, `spark.sql.legacy.parquet.datetimeRebaseModeInWrite/Read` are set to `EXCEPTION` by default.
### Why are the changes needed?
1. Parquet spec says that INT96 must be stored as Julian days (see https://github.com/apache/parquet-format/pull/49). This doesn't mean that a reader ( or a writer) is based on the Julian calendar. So, rebasing from Proleptic Gregorian to Julian calendar can be not needed.
2. Rebasing from/to Julian calendar can loose information because dates in one calendar don't exist in another one. Like 1582-10-04..1582-10-15 exist in Proleptic Gregorian calendar but not in the hybrid calendar (Julian + Gregorian), and visa versa, Julian date 1000-02-29 doesn't exist in Proleptic Gregorian calendar. We should allow users to save timestamps without loosing such dates (rebasing shifts such dates to the next valid date).
3. It would also make Spark compatible with other systems such as Impala and newer versions of Hive that write proleptic Gregorian based INT96 timestamps.
### Does this PR introduce _any_ user-facing change?
It can when `spark.sql.legacy.parquet.int96RebaseModeInWrite` is set non-default value `LEGACY`.
### How was this patch tested?
- Added a test to check the metadata key `org.apache.spark.int96NoRebase`
- By `ParquetIOSuite`
Closes#30056 from MaxGekk/parquet-rebase-int96.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Currently, actual non-dynamic partition pruning is executed in the optimizer phase (PruneFileSourcePartitions) if an input relation has a catalog file index. The current code assumes the same partition filters are generated again in FileSourceStrategy and passed into FileSourceScanExec. FileSourceScanExec uses the partition filters when listing files, but these non-dynamic partition filters do nothing because unnecessary partitions are already pruned in advance, so the filters are mainly used for explain output in this case. If a WHERE clause has DNF-ed predicates, FileSourceStrategy cannot extract the same filters with PruneFileSourcePartitions and then PartitionFilters is not shown in explain output.
This patch proposes to extract partition filters in FileSourceStrategy and HiveStrategy with `extractPredicatesWithinOutputSet` added in https://github.com/apache/spark/pull/29101/files#diff-6be42cfa3c62a7536b1eb1d6447c073c again, then It will show the partially pushed down partition filter in explain().
### Why are the changes needed?
without the patch, the explained plan is inconsistent with what is actually executed
<b>without the change </b> the explained plan of `"SELECT * FROM t WHERE p = '1' OR (p = '2' AND i = 1)"` for datasource and hive tables are like the following respectively (missing pushed down partition filters)
```
== Physical Plan ==
*(1) Filter ((p#21 = 1) OR ((p#21 = 2) AND (i#20 = 1)))
+- *(1) ColumnarToRow
+- FileScan parquet default.t[i#20,p#21] Batched: true, DataFilters: [], Format: Parquet, Location: InMemoryFileIndex[file:/Users/nanzhu/code/spark/sql/hive/target/tmp/hive_execution_test_group/war..., PartitionFilters: [], PushedFilters: [], ReadSchema: struct<i:int>
```
```
== Physical Plan ==
*(1) Filter ((p#33 = 1) OR ((p#33 = 2) AND (i#32 = 1)))
+- Scan hive default.t [i#32, p#33], HiveTableRelation [`default`.`t`, org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe, Data Cols: [i#32], Partition Cols: [p#33], Pruned Partitions: [(p=1), (p=2)]]
```
<b> with change </b> the plan looks like (the actually executed partition filters are exhibited)
```
== Physical Plan ==
*(1) Filter ((p#21 = 1) OR ((p#21 = 2) AND (i#20 = 1)))
+- *(1) ColumnarToRow
+- FileScan parquet default.t[i#20,p#21] Batched: true, DataFilters: [], Format: Parquet, Location: InMemoryFileIndex[file:/Users/nanzhu/code/spark/sql/hive/target/tmp/hive_execution_test_group/war..., PartitionFilters: [((p#21 = 1) OR (p#21 = 2))], PushedFilters: [], ReadSchema: struct<i:int>
```
```
== Physical Plan ==
*(1) Filter ((p#37 = 1) OR ((p#37 = 2) AND (i#36 = 1)))
+- Scan hive default.t [i#36, p#37], HiveTableRelation [`default`.`t`, org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe, Data Cols: [i#36], Partition Cols: [p#37], Pruned Partitions: [(p=1), (p=2)]], [((p#37 = 1) OR (p#37 = 2))]
```
### Does this PR introduce _any_ user-facing change
no
### How was this patch tested?
Unit test.
Closes#29831 from CodingCat/SPARK-32351.
Lead-authored-by: Nan Zhu <nanzhu@uber.com>
Co-authored-by: Nan Zhu <CodingCat@users.noreply.github.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Add MyPy to the CI. Once this is installed on the CI: https://issues.apache.org/jira/browse/SPARK-32797?jql=project%20%3D%20SPARK%20AND%20text%20~%20mypy this wil automatically check the types.
### Why are the changes needed?
We should check if the types are still correct on the CI.
```
MacBook-Pro-van-Fokko:spark fokkodriesprong$ ./dev/lint-python
starting python compilation test...
python compilation succeeded.
starting pycodestyle test...
pycodestyle checks passed.
starting flake8 test...
flake8 checks passed.
starting mypy test...
mypy checks passed.
The sphinx-build command was not found. Skipping Sphinx build for now.
all lint-python tests passed!
```
### Does this PR introduce _any_ user-facing change?
No :)
### How was this patch tested?
By running `./dev/lint-python` locally.
Closes#30088 from Fokko/SPARK-17333.
Authored-by: Fokko Driesprong <fokko@apache.org>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This patch proposes to add more optimization to `UpdateFields` expression chain. And optimize `UpdateFields` early in analysis phase.
### Why are the changes needed?
`UpdateFields` can manipulate complex nested data, but using `UpdateFields` can easily create inefficient expression chain. We should optimize it further.
Because when manipulating deeply nested schema, the `UpdateFields` expression tree could be too complex to analyze, this change optimizes `UpdateFields` early in analysis phase.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Unit test.
Closes#29812 from viirya/SPARK-32941.
Authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
1. Add the common trait `CommonFileDataSourceSuite` with tests that can be executed for all built-in file-based datasources.
2. Add a test `CommonFileDataSourceSuite` to check that datasource options are propagated to underlying file systems as Hadoop configs.
3. Mix `CommonFileDataSourceSuite` to `AvroSuite`, `OrcSourceSuite`, `TextSuite`, `JsonSuite`, CSVSuite` and to `ParquetFileFormatSuite`.
4. Remove duplicated tests from `AvroSuite` and from `OrcSourceSuite`.
### Why are the changes needed?
To improve test coverage and test all built-in file-based datasources.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
By running the affected test suites.
Closes#30067 from MaxGekk/ds-options-common-test.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
In [SPARK-33139](https://github.com/apache/spark/pull/30042), I was using reflect "Class.forName" in python code to invoke method in SparkSession which is not recommended. using getattr to access "SparkSession$.Module$" instead.
### Why are the changes needed?
Code refine.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Existing tests.
Closes#30092 from leanken/leanken-SPARK-33139-followup.
Authored-by: xuewei.linxuewei <xuewei.linxuewei@alibaba-inc.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR aims to switch the default Hadoop profile from `hadoop2.7` to `hadoop3.2` in `dev/run-tests.py` when it's running in local or GitHub Action environments.
### Why are the changes needed?
The default Hadoop version is 3.2. We had better be consistent.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Manually.
**BEFORE**
```
% dev/run-tests.py
Cannot install SparkR as R was not found in PATH
[info] Using build tool sbt with Hadoop profile hadoop2.7 and Hive profile hive2.3 under environment local
```
**AFTER**
```
% dev/run-tests.py
Cannot install SparkR as R was not found in PATH
[info] Using build tool sbt with Hadoop profile hadoop3.2 and Hive profile hive2.3 under environment local
```
Closes#30090 from williamhyun/SPARK-33179.
Authored-by: William Hyun <williamhyun3@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR aims to ignore GitHub only changes in Amplab Jenkins build.
### Why are the changes needed?
This will save server resources.
### Does this PR introduce _any_ user-facing change?
No, this is a dev-only change.
### How was this patch tested?
Manually. I used the following doctest during testing and removed it at the clean-up.
E2E tests:
```
cd dev
cat test.py
```
```python
import importlib
runtests = importlib.import_module("run-tests")
print([x.name for x in runtests.determine_modules_for_files([".github/workflows/build_and_test.yml"])])
```
```python
$ GITHUB_ACTIONS=1 python test.py
['root']
$ python test.py
[]
```
Unittests:
```bash
$ GITHUN_ACTIONS=1 python3 -m doctest dev/run-tests.py
$ python3 -m doctest dev/run-tests.py
```
Closes#30020 from williamhyun/SPARK-33123.
Lead-authored-by: William Hyun <williamhyun3@gmail.com>
Co-authored-by: Hyukjin Kwon <gurwls223@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Improve error message on reading unexpected directory
### Why are the changes needed?
Improve error message on reading unexpected directory
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Ut
Closes#30027 from AngersZhuuuu/SPARK-32069.
Authored-by: angerszhu <angers.zhu@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Mark `CollectList` and `CollectSet` as non-nullable.
### Why are the changes needed?
`CollectList` and `CollectSet` 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.
Closes#30087 from tanelk/SPARK-33177_collect.
Authored-by: tanel.kiis@gmail.com <tanel.kiis@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR aims to detect duplicate `mountPath`s and stop the job.
### Why are the changes needed?
If there is a conflict on `mountPath`, the pod is created and repeats the following error messages and keeps running. Spark job should not keep running and wasting the cluster resources. We had better fail at Spark side.
```
$ k get pod -l 'spark-role in (driver,executor)'
NAME READY STATUS RESTARTS AGE
tpcds 1/1 Running 0 33m
```
```
20/10/18 05:09:26 WARN ExecutorPodsSnapshotsStoreImpl: Exception when notifying snapshot subscriber.
io.fabric8.kubernetes.client.KubernetesClientException: Failure executing: POST at: ...
Message: Pod "tpcds-exec-1" is invalid: spec.containers[0].volumeMounts[1].mountPath:
Invalid value: "/data1": must be unique.
...
```
**AFTER THIS PR**
The job will stop with the following error message instead of keeping running.
```
20/10/18 06:58:45 ERROR ExecutorPodsSnapshotsStoreImpl: Going to stop due to IllegalArgumentException
java.lang.IllegalArgumentException: requirement failed: Found duplicated mountPath: `/data1`
```
### Does this PR introduce _any_ user-facing change?
Yes, but this is a bug fix.
### How was this patch tested?
Pass the CI with the newly added test case.
Closes#30084 from dongjoon-hyun/SPARK-33175-2.
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 proposes to remove an obsolete comment about adding the `sbt-dependency-graph` back in SBT plugins.
### Why are the changes needed?
sbt-dependency-graph is now built-in from SBT 1.4.0, see https://github.com/sbt/sbt/releases/tag/v1.4.0.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Manually tested `./build/sbt dependencyTree`.
Closes#30085 from HyukjinKwon/SPARK-33109.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This PR aims to use `openjdk:11-jre-slim` as default in K8s Dockerfile.
### Why are the changes needed?
Although Apache Spark supports both Java8/Java11, there is a difference.
1. Java8-built distribution can run both Java8/Java11
2. Java11-built distribution can run on Java11, but not Java8.
In short, we had better use Java11 in Dockerfile to embrace both cases without any issues.
### Does this PR introduce _any_ user-facing change?
Yes. This will remove the change of user frustration when they build with JDK11 and build the image without overriding Java base image.
### How was this patch tested?
Pass the K8s IT.
Closes#30083 from dongjoon-hyun/SPARK-33176.
Authored-by: Dongjoon Hyun <dongjoon@apache.org>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This pull request changes the link to the pickle module page from https://docs.python.org/2/library/pickle.html to https://docs.python.org/3/library/pickle.html in RDD Programming Guide.
### Why are the changes needed?
Since Python 2 is no longer supported and it is preferable to refer to the pickle module page of Python 3.
### Does this PR introduce _any_ user-facing change?
Yes.
Before: the `Pickle` link's destination page was https://docs.python.org/2/library/pickle.html
After: the `Pickle` link's destination page is https://docs.python.org/3/library/pickle.html
### How was this patch tested?
By building the documentation site and check the link's destination page is changed correctly in the local environment.
Closes#30081 from kjmrknsn/docs-fix-pickle-link.
Authored-by: Keiji Yoshida <kjmrknsn@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This pull request changes the `sep` parameter's value from `:` to `;` in the example of `examples/src/main/python/sql/datasource.py`. This code snippet is shown on the Spark SQL Guide documentation. The `sep` parameter's value should be `;` since the data in https://github.com/apache/spark/blob/master/examples/src/main/resources/people.csv is separated by `;`.
### Why are the changes needed?
To fix the example code so that it can be executed properly.
### Does this PR introduce _any_ user-facing change?
Yes.
This code snippet is shown on the Spark SQL Guide documentation: https://spark.apache.org/docs/latest/sql-data-sources-load-save-functions.html#manually-specifying-options
### How was this patch tested?
By building the documentation and checking the Spark SQL Guide documentation manually in the local environment.
Closes#30082 from kjmrknsn/fix-example-python-datasource.
Authored-by: Keiji Yoshida <kjmrknsn@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This patch proposes to add a config we can control fast-fail behavior in FileFormatWriter and set it false by default.
### Why are the changes needed?
In SPARK-29649, we catch `FileAlreadyExistsException` in `FileFormatWriter` and fail fast for the task set to prevent task retry.
Due to latest discussion, it is important to be able to keep original behavior that is to retry tasks even `FileAlreadyExistsException` is thrown, because `FileAlreadyExistsException` could be recoverable in some cases.
We are going to add a config we can control this behavior and set it false for fast-fail by default.
### Does this PR introduce _any_ user-facing change?
Yes. By default the task in FileFormatWriter will retry even if `FileAlreadyExistsException` is thrown. This is the behavior before Spark 3.0. User can control fast-fail behavior by enabling it.
### How was this patch tested?
Unit test.
Closes#30073 from viirya/SPARK-33170.
Authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This is a minor change to update structured-streaming-programming-guide and typos in code.
### Why are the changes needed?
Keep the user-facing document correct and updated.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Unit tests.
Closes#30074 from viirya/ss-minor.
Authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This PR aims to use `eventually` to fix the flakiness of the test case `SPARK-33088: executor failed tasks trigger plugin calls`.
### Why are the changes needed?
The test case checks like the following.
```scala
assert(TestSparkPlugin.executorPlugin.numOnTaskStart == 2)
assert(TestSparkPlugin.executorPlugin.numOnTaskSucceeded == 0)
assert(TestSparkPlugin.executorPlugin.numOnTaskFailed == 2)
```
Although first and second passed, the third can fail.
- https://amplab.cs.berkeley.edu/jenkins/view/Spark%20QA%20Test%20(Dashboard)/job/spark-master-test-maven-hadoop-3.2-hive-2.3-jdk-11/lastCompletedBuild/testReport/org.apache.spark.internal.plugin/PluginContainerSuite/SPARK_33088__executor_failed_tasks_trigger_plugin_calls/
- https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/129919/testReport/
```
sbt.ForkMain$ForkError: org.scalatest.exceptions.TestFailedException: 1 did not equal 2
at org.scalatest.Assertions.newAssertionFailedException(Assertions.scala:472)
at org.scalatest.Assertions.newAssertionFailedException$(Assertions.scala:471)
at org.scalatest.Assertions$.newAssertionFailedException(Assertions.scala:1231)
at org.scalatest.Assertions$AssertionsHelper.macroAssert(Assertions.scala:1295)
at org.apache.spark.internal.plugin.PluginContainerSuite.$anonfun$new$8(PluginContainerSuite.scala:161)
```
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
This only improves the robustness.
Closes#30072 from dongjoon-hyun/SPARK-33173.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Upgrade sbt to release 1.4.0
### Why are the changes needed?
Bring built-in `dependencyTree` instead of removed `sbt-dependency-graph` plugin, that doesn't work with sbt used in build.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Should pass all the tests.
Closes#30070 from gemelen/feature/sbt-1.4.
Authored-by: Denis Pyshev <git@gemelen.net>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
SPARK-29358 added support for `unionByName` to work when the two datasets didn't necessarily have the same schema, but it does not work with nested columns like structs. This patch adds the support to work with struct columns.
The behavior before this PR:
```scala
scala> val df1 = spark.range(1).selectExpr("id c0", "named_struct('c', id + 1, 'b', id + 2, 'a', id + 3) c1")
scala> val df2 = spark.range(1).selectExpr("id c0", "named_struct('c', id + 1, 'b', id + 2) c1")
scala> df1.unionByName(df2, true).printSchema
org.apache.spark.sql.AnalysisException: Union can only be performed on tables with the compatible column types. struct<c:bigint,b:bigint> <> struct<c:bigint,b:bigint,a:bigint> at the second column of the second table;;
'Union false, false
:- Project [id#0L AS c0#2L, named_struct(c, (id#0L + cast(1 as bigint)), b, (id#0L + cast(2 as bigint)), a, (id#0L + cast(3 as bigint))) AS c1#3]
: +- Range (0, 1, step=1, splits=Some(12))
+- Project [c0#8L, c1#9]
+- Project [id#6L AS c0#8L, named_struct(c, (id#6L + cast(1 as bigint)), b, (id#6L + cast(2 as bigint))) AS c1#9]
+- Range (0, 1, step=1, splits=Some(12))
```
The behavior after this PR:
```scala
scala> df1.unionByName(df2, true).printSchema
root
|-- c0: long (nullable = false)
|-- c1: struct (nullable = false)
| |-- a: long (nullable = true)
| |-- b: long (nullable = false)
| |-- c: long (nullable = false)
scala> df1.unionByName(df2, true).show()
+---+-------------+
| c0| c1|
+---+-------------+
| 0| {3, 2, 1}|
| 0|{ null, 2, 1}|
+---+-------------+
```
### Why are the changes needed?
The `allowMissingColumns` of `unionByName` is a feature allowing merging different schema from two datasets when unioning them together. Nested column support makes the feature more general and flexible for usage.
### Does this PR introduce _any_ user-facing change?
Yes, after this change users can union two datasets with different schema with different structs.
### How was this patch tested?
Unit tests.
Closes#29587 from viirya/SPARK-32376.
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?
If a block is removed between discovery to transfer fo the block, we short circuit that block and remove it from the list to transfer and increment the transferred blocks. This is complicated since both RPC errors and local read errors may be reported with the same exception class.
### Why are the changes needed?
Slow shuffle refreshes could waste time when decommissioning has already finished. Decommissioning might avoid transferring some some blocks to an otherwise live host which is marked as "full" if a deleted block fails to transfer to that host.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
New unit and integration tests.
Closes#30046 from holdenk/handle-cleaned-shuffles-during0migration.
Authored-by: Holden Karau <hkarau@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
The function `binaryToSQLTimestamp()` is used by Parquet Vectorized reader. Parquet MR reader has similar code for de-serialization of INT96 timestamps. In this PR, I propose to de-duplicate code and re-use `binaryToSQLTimestamp()`.
### Why are the changes needed?
This should improve maintenance, and should allow to avoid errors while changing Vectorized and regular parquet readers.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
By existing test suites, for instance `ParquetIOSuite`.
Closes#30069 from MaxGekk/int96-common-serde.
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 PR aims to mark ParquetV1FilterSuite and ParquetV2FilterSuite as `ExtendedSQLTest`.
- ParquetV1FilterSuite/ParquetV2FilterSuite
- ParquetV1SchemaPruningSuite/ParquetV2SchemaPruningSuite
### Why are the changes needed?
Currently, `sql - other tests` is the longest job. This PR will move the above tests to `sql - slow tests` job.
**BEFORE**
- https://github.com/apache/spark/runs/1264150802 (1 hour 37 minutes)
**AFTER**
- https://github.com/apache/spark/pull/30068/checks?check_run_id=1265879896 (1 hour 21 minutes)
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Pass the Github Action with the reduced time.
Closes#30068 from dongjoon-hyun/MOVE3.
Lead-authored-by: Dongjoon Hyun <dongjoon@apache.org>
Co-authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Hive's `hive-service-rpc` module started since hive-2.1.0 and it contains only the thrift IDL file and the code generated by it.
Removing the inlined code will help maintain and upgrade builtin hive versions
### Why are the changes needed?
to simply the code.
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
passing CI
Closes#30055 from yaooqinn/SPARK-33159.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
### What changes were proposed in this pull request?
In Execution web page, when `Succeeded Job`(or Failed Jobs) has many child url elements,they will extend over the edge of the page.
### Why are the changes needed?
To make the page more friendly.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Munual test result shows as below:
Closes#30035 from akiyamaneko/sql_execution_job_overflow.
Authored-by: neko <echohlne@gmail.com>
Signed-off-by: Gengliang Wang <gengliang.wang@databricks.com>
### What changes were proposed in this pull request?
Correct the resolution of having clause.
### Why are the changes needed?
Grouping sets construct new aggregate lost the qualified name of grouping expression. Here is a example:
```
-- Works resolved by `ResolveReferences`
select c1 from values (1) as t1(c1) group by grouping sets(t1.c1) having c1 = 1
-- Works because of the extra expression c1
select c1 as c2 from values (1) as t1(c1) group by grouping sets(t1.c1) having t1.c1 = 1
-- Failed
select c1 from values (1) as t1(c1) group by grouping sets(t1.c1) having t1.c1 = 1
```
It wroks with `Aggregate` without grouping sets through `ResolveReferences`, but Grouping sets not works since the exprId has been changed.
### Does this PR introduce _any_ user-facing change?
Yes, bug fix.
### How was this patch tested?
add test.
Closes#30029 from ulysses-you/SPARK-33131.
Authored-by: ulysses <youxiduo@weidian.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
The current `Lead`/`Lag` extends `OffsetWindowFunction`. `OffsetWindowFunction` contains field `direction` and use `direction` to calculates the `boundary`.
We can use single literal expression unify the two properties.
For example:
3 means `direction` is Asc and `boundary` is 3.
-3 means `direction` is Desc and `boundary` is -3.
### Why are the changes needed?
Improve the current implement of `Lead`/`Lag`.
### Does this PR introduce _any_ user-facing change?
'No'.
### How was this patch tested?
Jenkins test.
Closes#30023 from beliefer/SPARK-33126.
Lead-authored-by: gengjiaan <gengjiaan@360.cn>
Co-authored-by: beliefer <beliefer@163.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR is a sub-task of [SPARK-33138](https://issues.apache.org/jira/browse/SPARK-33138). In order to make SQLConf.get reliable and stable, we need to make sure user can't pollute the SQLConf and SparkSession Context via calling setActiveSession and clearActiveSession.
Change of the PR:
* add legacy config spark.sql.legacy.allowModifyActiveSession to fallback to old behavior if user do need to call these two API.
* by default, if user call these two API, it will throw exception
* add extra two internal and private API setActiveSessionInternal and clearActiveSessionInternal for current internal usage
* change all internal reference to new internal API except for SQLContext.setActive and SQLContext.clearActive
### Why are the changes needed?
Make SQLConf.get reliable and stable.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
* Add UT in SparkSessionBuilderSuite to test the legacy config
* Existing test
Closes#30042 from leanken/leanken-SPARK-33139.
Authored-by: xuewei.linxuewei <xuewei.linxuewei@alibaba-inc.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR proposes to use `scala.Predef.assert` instead of `org.scalatest.Assertions.assert` removed at https://github.com/apache/spark/pull/30064
### Why are the changes needed?
Just to keep the same behaviour.
### Does this PR introduce _any_ user-facing change?
No, dev-only
### How was this patch tested?
Recover the existing asserts.
Closes#30065 from HyukjinKwon/SPARK-33165.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Proposing a new set of APIs for ExecutorPlugins, to provide callbacks invoked at the start and end of each task of a job. Not very opinionated on the shape of the API, tried to be as minimal as possible for now.
### Why are the changes needed?
Changes described in detail on [SPARK-33088](https://issues.apache.org/jira/browse/SPARK-33088), but mostly this boils down to:
1. This feature was considered when the ExecutorPlugin API was initially introduced in #21923, but never implemented.
2. The use-case which **requires** this feature is to propagate tracing information from the driver to the executor, such that calls from the same job can all be traced.
a. Tracing frameworks usually are setup in thread locals, therefore it's important for the setup to happen in the same thread which runs the tasks.
b. Executors can be for multiple jobs, therefore it's not sufficient to set tracing information at executor startup time -- it needs to happen every time a task starts or ends.
### Does this PR introduce _any_ user-facing change?
No. This PR introduces new features for future developers to use.
### How was this patch tested?
Unit tests on `PluginContainerSuite`.
Closes#29977 from fsamuel-bs/SPARK-33088.
Authored-by: Samuel Souza <ssouza@palantir.com>
Signed-off-by: Mridul Muralidharan <mridul<at>gmail.com>
### What changes were proposed in this pull request?
This PR proposes to remove `assert` from `Benchmark` for making it easier to run benchmark codes via `spark-submit`.
### Why are the changes needed?
Since the current `Benchmark` (`master` and `branch-3.0`) has `assert`, we need to pass the proper jars of `scalatest` and `scalactic`;
- scalatest-core_2.12-3.2.0.jar
- scalatest-compatible-3.2.0.jar
- scalactic_2.12-3.0.jar
```
./bin/spark-submit --jars scalatest-core_2.12-3.2.0.jar,scalatest-compatible-3.2.0.jar,scalactic_2.12-3.0.jar,./sql/catalyst/target/spark-catalyst_2.12-3.1.0-SNAPSHOT-tests.jar,./core/target/spark-core_2.12-3.1.0-SNAPSHOT-tests.jar --class org.apache.spark.sql.execution.benchmark.TPCDSQueryBenchmark ./sql/core/target/spark-sql_2.12-3.1.0-SNAPSHOT-tests.jar --data-location /tmp/tpcds-sf1
```
This update can make developers submit benchmark codes without these dependencies;
```
./bin/spark-submit --jars ./sql/catalyst/target/spark-catalyst_2.12-3.1.0-SNAPSHOT-tests.jar,./core/target/spark-core_2.12-3.1.0-SNAPSHOT-tests.jar --class org.apache.spark.sql.execution.benchmark.TPCDSQueryBenchmark ./sql/core/target/spark-sql_2.12-3.1.0-SNAPSHOT-tests.jar --data-location /tmp/tpcds-sf1
```
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Manually checked.
Closes#30064 from maropu/RemoveDepInBenchmark.
Authored-by: Takeshi Yamamuro <yamamuro@apache.org>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Add case sensitivity tests for column resolution in ALTER TABLE
### Why are the changes needed?
To make sure `spark.sql.caseSensitive` works for `ResolveAlterTableChanges`
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
new test
Closes#30063 from huaxingao/caseSensitivity.
Authored-by: Huaxin Gao <huaxing@us.ibm.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Added a couple tests to `AvroSuite` and to `ParquetIOSuite` to check that the metadata key 'org.apache.spark.legacyDateTime' is written correctly depending on the SQL configs:
- spark.sql.legacy.avro.datetimeRebaseModeInWrite
- spark.sql.legacy.parquet.datetimeRebaseModeInWrite
This is a follow up https://github.com/apache/spark/pull/28137.
### Why are the changes needed?
1. To improve test coverage
2. To make sure that the metadata key is actually saved to Avro/Parquet files
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
By running the added tests:
```
$ build/sbt "testOnly org.apache.spark.sql.execution.datasources.parquet.ParquetIOSuite"
$ build/sbt "avro/test:testOnly org.apache.spark.sql.avro.AvroV1Suite"
$ build/sbt "avro/test:testOnly org.apache.spark.sql.avro.AvroV2Suite"
```
Closes#30061 from MaxGekk/parquet-test-metakey.
Authored-by: Max Gekk <max.gekk@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Change `df.groupBy.agg()` to `df.groupBy().agg()` in the docstring of `agg()`
### Why are the changes needed?
Fix typo in a docstring
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
No
Closes#30060 from ChuliangXiao/patch-1.
Authored-by: Chuliang Xiao <ChuliangX@gmail.com>
Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
### What changes were proposed in this pull request?
Current solution in build file to enable build failure on compilation warnings with exclusion of deprecation ones is not portable after SBT version 1.3.13 (build import fails with compilation error with SBT 1.4) and could be replaced with more robust and maintainable, especially since Scala 2.13.2 with similar built-in functionality.
Additionally, warnings were fixed to pass the build, with as few changes as possible:
warnings in 2.12 compilation fixed in code,
warnings in 2.13 compilation covered by configuration to be addressed separately
### Why are the changes needed?
Unblocks upgrade to SBT after 1.3.13.
Enhances build file maintainability.
Allows fine tune of warnings configuration in scope of Scala 2.13 compilation.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
`build/sbt`'s `compile` and `Test/compile` for both Scala 2.12 and 2.13 profiles.
Closes#29995 from gemelen/feature/warnings-reporter.
Authored-by: Denis Pyshev <git@gemelen.net>
Signed-off-by: Sean Owen <srowen@gmail.com>
### What changes were proposed in this pull request?
This proposes to add a config for json expression optimization.
### Why are the changes needed?
For the new Json expression optimization rules, it is safer if we can disable it using SQL config.
### Does this PR introduce _any_ user-facing change?
Yes, users can disable json expression optimization rule.
### How was this patch tested?
Unit test
Closes#30047 from viirya/SPARK-33078.
Authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
### What changes were proposed in this pull request?
This is the first patch for SPIP SPARK-30602 for push-based shuffle.
Summary of changes:
* Introduce new API in ExternalBlockStoreClient to push blocks to a remote shuffle service.
* Leveraging the streaming upload functionality in SPARK-6237, it also enables the ExternalBlockHandler to delegate the handling of block push requests to MergedShuffleFileManager.
* Propose the API for MergedShuffleFileManager, where the core logic on the shuffle service side to handle block push requests is defined. The actual implementation of this API is deferred into a later RB to restrict the size of this PR.
* Introduce OneForOneBlockPusher to enable pushing blocks to remote shuffle services in shuffle RPC layer.
* New protocols in shuffle RPC layer to support the functionalities.
### Why are the changes needed?
Refer to the SPIP in SPARK-30602
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Added unit tests.
The reference PR with the consolidated changes covering the complete implementation is also provided in SPARK-30602.
We have already verified the functionality and the improved performance as documented in the SPIP doc.
Lead-authored-by: Min Shen <mshenlinkedin.com>
Co-authored-by: Chandni Singh <chsinghlinkedin.com>
Co-authored-by: Ye Zhou <yezhoulinkedin.com>
Closes#29855 from Victsm/SPARK-32915.
Lead-authored-by: Min Shen <mshen@linkedin.com>
Co-authored-by: Chandni Singh <chsingh@linkedin.com>
Co-authored-by: Ye Zhou <yezhou@linkedin.com>
Co-authored-by: Chandni Singh <singh.chandni@gmail.com>
Co-authored-by: Min Shen <victor.nju@gmail.com>
Signed-off-by: Mridul Muralidharan <mridul<at>gmail.com>
### What changes were proposed in this pull request?
This PR proposes to install `scipy` as well in PyPy. It will test several ML specific test cases in PyPy as well. For example, 31a16fbb40/python/pyspark/mllib/tests/test_linalg.py (L487)
It was not installed when GitHub Actions build was added because it failed to install for an unknown reason. Seems like it's fixed in the latest scipy.
### Why are the changes needed?
To improve test coverage.
### Does this PR introduce _any_ user-facing change?
No, dev-only.
### How was this patch tested?
GitHub Actions build in this PR will test it out.
Closes#30054 from HyukjinKwon/SPARK-32247.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
I currently have unquoted column names in alter table, e.g. ```ALTER TABLE "test"."alt_table" DROP COLUMN c1```
should change to quoted column name ```ALTER TABLE "test"."alt_table" DROP COLUMN "c1"```
### Why are the changes needed?
We should always use quoted identifiers in JDBC SQLs, e.g. ```CREATE TABLE "test"."abc" ("col" INTEGER ) ``` or ```INSERT INTO "test"."abc" ("col") VALUES (?)```. Using unquoted column name in alterTable causes problems, for example:
```
sql("CREATE TABLE h2.test.alt_table (c1 INTEGER, c2 INTEGER) USING _")
sql("ALTER TABLE h2.test.alt_table DROP COLUMN c1")
org.apache.spark.sql.AnalysisException: Failed table altering: test.alt_table;
......
Caused by: org.h2.jdbc.JdbcSQLException: Column "C1" not found; SQL statement:
ALTER TABLE "test"."alt_table" DROP COLUMN c1 [42122-195]
```
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existing tests
Closes#30041 from huaxingao/alter_table_followup.
Authored-by: Huaxin Gao <huaxing@us.ibm.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
SPARK-32926 added a build test to GitHub Action for Scala 2.13 but it's only with Maven.
As SPARK-32873 reported, some compilation error happens only with SBT so I think we need to add another build test to GitHub Action for SBT.
Unfortunately, we don't have abundant resources for GitHub Actions so instead of just adding the new SBT job, let's replace the existing Maven job with the new SBT job for Scala 2.13.
### Why are the changes needed?
To ensure build test passes even with SBT for Scala 2.13.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
GitHub Actions' job.
Closes#29958 from sarutak/add-sbt-job-for-scala-2.13.
Authored-by: Kousuke Saruta <sarutak@oss.nttdata.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR aims to upgrade `Github Action` runner image from `Ubuntu 18.04 (LTS)` to `Ubuntu 20.04 (LTS)`.
### Why are the changes needed?
`ubuntu-latest` in `GitHub Action` is still `Ubuntu 18.04 (LTS)`.
- https://github.com/actions/virtual-environments#available-environments
This upgrade will help Apache Spark 3.1+ preparation for vote and release on the latest OS.
This is tested here.
- https://github.com/dongjoon-hyun/spark/pull/36
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Pass the `Github Action` in this PR.
Closes#30050 from dongjoon-hyun/ubuntu_20.04.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This PR makes `spark.kubernetes.pyspark.pythonVersion` allow only `3`. In other words, it will reject `2` for `Python 2`.
- [x] Configuration description and check is updated.
- [x] Documentation is updated
- [x] Unit test cases are updated.
- [x] Docker image script is updated.
### Why are the changes needed?
After SPARK-32138, Apache Spark 3.1 dropped Python 2 support.
### Does this PR introduce _any_ user-facing change?
Yes, but Python 2 support is already dropped officially.
### How was this patch tested?
Pass the CI.
Closes#30049 from dongjoon-hyun/SPARK-DROP-PYTHON2.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Do not use local shuffle reader at final stage if the root node is write command.
### Why are the changes needed?
Users usually repartition with partition column on dynamic partition overwrite. AQE could break it by removing physical shuffle with local shuffle reader. That could lead to a large number of output files, even exceeding the file system limit.
### Does this PR introduce _any_ user-facing change?
Yes.
### How was this patch tested?
Add test.
Closes#29797 from manuzhang/spark-32932.
Authored-by: manuzhang <owenzhang1990@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR aims to ignore Apache Spark 2.4.x distribution in HiveExternalCatalogVersionsSuite if Python version is 3.8 or 3.9.
### Why are the changes needed?
Currently, `HiveExternalCatalogVersionsSuite` is broken on the latest OS like `Ubuntu 20.04` because its default Python version is 3.8. PySpark 2.4.x doesn't work on Python 3.8 due to SPARK-29536.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Manually.
```
$ python3 --version
Python 3.8.5
$ build/sbt "hive/testOnly *.HiveExternalCatalogVersionsSuite"
...
[info] All tests passed.
[info] Passed: Total 1, Failed 0, Errors 0, Passed 1
```
Closes#30044 from dongjoon-hyun/SPARK-33153.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Adds an additional check for non-fatal errors when attempting to add a new entry to the history server application listing.
### Why are the changes needed?
A bad rolling event log folder (missing appstatus file or no log files) would cause no applications to be loaded by the Spark history server. Figuring out why invalid event log folders are created in the first place will be addressed in separate issues, this just lets the history server skip the invalid folder and successfully load all the valid applications.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
New UT
Closes#30037 from Kimahriman/bug/rolling-log-crashing-history.
Authored-by: Adam Binford <adam.binford@radiantsolutions.com>
Signed-off-by: Jungtaek Lim (HeartSaVioR) <kabhwan.opensource@gmail.com>
### What changes were proposed in this pull request?
Follow the convention and rename the metrics `numRows` to `numOutputRows`
### Why are the changes needed?
`FilterExec`, `HashAggregateExec`, etc. all use `numOutputRows`
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
existing tests
Closes#30039 from cloud-fan/minor.
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 proposes to fix a bug on calling `DataType.equalsIgnoreCompatibleNullability` with mistakenly swapped parameters in `V2WriteCommand.outputResolved`. The order of parameters for `DataType.equalsIgnoreCompatibleNullability` are `from` and `to`, which says that the right order of matching variables are `inAttr` and `outAttr`.
### Why are the changes needed?
Spark throws AnalysisException due to unresolved operator in v2 write, while the operator is unresolved due to a bug that parameters to call `DataType.equalsIgnoreCompatibleNullability` in `outputResolved` have been swapped.
### Does this PR introduce _any_ user-facing change?
Yes, end users no longer suffer on unresolved operator in v2 write if they're trying to write dataframe containing non-nullable complex types against table matching complex types as nullable.
### How was this patch tested?
New UT added.
Closes#30033 from HeartSaVioR/SPARK-33136.
Authored-by: Jungtaek Lim (HeartSaVioR) <kabhwan.opensource@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
