### What changes were proposed in this pull request?
additional links to other classes in python documentation
### Why are the changes needed?
python docstring syntax wasn't fully used everywhere
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Documentation change only
Closes#33440 from dominikgehl/feature/python-docstrings.
Authored-by: Dominik Gehl <dog@open.ch>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit 701756ac95)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Trying to adjust build memory settings and serial execution to re-enable GA.
### Why are the changes needed?
GA tests are failed recently due to return code 137. We need to adjust build settings to make GA work.
### Does this PR introduce _any_ user-facing change?
No, dev only.
### How was this patch tested?
GA
Closes#33447 from viirya/test-ga.
Lead-authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Co-authored-by: Hyukjin Kwon <gurwls223@gmail.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit fd36ed4550)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Add `remove_unused_categories` to `CategoricalAccessor` and `CategoricalIndex`.
### Why are the changes needed?
We should implement `remove_unused_categories` in `CategoricalAccessor` and `CategoricalIndex`.
### Does this PR introduce _any_ user-facing change?
Yes, users will be able to use `remove_unused_categories`.
### How was this patch tested?
Added some tests.
Closes#33485 from ueshin/issues/SPARK-36261/remove_unused_categories.
Authored-by: Takuya UESHIN <ueshin@databricks.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit 2fe12a7520)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR proposes to set the lowerbound of mypy version to use in the testing script.
### Why are the changes needed?
https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder/141519/console
```
python/pyspark/mllib/tree.pyi:29: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/tree.pyi:38: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/feature.pyi:34: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/feature.pyi:42: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/feature.pyi:48: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/feature.pyi:54: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/feature.pyi:76: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/feature.pyi:124: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/feature.pyi:165: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/clustering.pyi:45: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/clustering.pyi:72: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/classification.pyi:39: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
python/pyspark/mllib/classification.pyi:52: error: Overloaded function signatures 1 and 2 overlap with incompatible return types
Found 13 errors in 4 files (checked 314 source files)
1
```
Jenkins installed mypy at SPARK-32797 but seems the version installed is not same as GIthub Actions.
It seems difficult to make the codebase compatible with multiple mypy versions. Therefore, this PR sets the lowerbound.
### Does this PR introduce _any_ user-facing change?
No, dev-only.
### How was this patch tested?
Jenkins job in this PR should test it out.
Also manually tested:
Without mypy:
```
...
flake8 checks passed.
The mypy command was not found. Skipping for now.
```
With mypy 0.812:
```
...
flake8 checks passed.
The minimum mypy version needs to be 0.910. Your current version is mypy 0.812. Skipping for now.
```
With mypy 0.910:
```
...
flake8 checks passed.
starting mypy test...
mypy checks passed.
all lint-python tests passed!
```
Closes#33487 from HyukjinKwon/SPARK-36268.
Authored-by: Hyukjin Kwon <gurwls223@apache.org>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit d6bc8cd681)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Improve bool, string, numeric DataTypeOps tests by avoiding joins.
Previously, bool, string, numeric DataTypeOps tests are conducted between two different Series.
After the PR, bool, string, numeric DataTypeOps tests should perform on a single DataFrame.
### Why are the changes needed?
A considerable number of DataTypeOps tests have operations on different Series, so joining is needed, which takes a long time.
We shall avoid joins for a shorter test duration.
The majority of joins happen in bool, string, numeric DataTypeOps tests, so we improve them first.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Unit tests.
Closes#33402 from xinrong-databricks/datatypeops_diffframe.
Authored-by: Xinrong Meng <xinrong.meng@databricks.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit 75fd1f5b82)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Use `Column.__getitem__` instead of `Column.getItem` to suppress warnings.
### Why are the changes needed?
In pandas API on Spark code base, there are some places using `Column.getItem` with `Column` object, but it shows a deprecation warning.
### Does this PR introduce _any_ user-facing change?
Yes, users won't see the warnings anymore.
- before
```py
>>> s = ps.Series(list("abbccc"), dtype="category")
>>> s.astype(str)
/path/to/spark/python/pyspark/sql/column.py:322: FutureWarning: A column as 'key' in getItem is deprecated as of Spark 3.0, and will not be supported in the future release. Use `column[key]` or `column.key` syntax instead.
warnings.warn(
0 a
1 b
2 b
3 c
4 c
5 c
dtype: object
```
- after
```py
>>> s = ps.Series(list("abbccc"), dtype="category")
>>> s.astype(str)
0 a
1 b
2 b
3 c
4 c
5 c
dtype: object
```
### How was this patch tested?
Existing tests.
Closes#33486 from ueshin/issues/SPARK-36265/getitem.
Authored-by: Takuya UESHIN <ueshin@databricks.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit a76a087f7f)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR aims to upgrade ZSTD-JNI to 1.5.0-4.
### Why are the changes needed?
ZSTD-JNI 1.5.0-3 has a packaging issue. 1.5.0-4 is recommended to be used instead.
- https://github.com/luben/zstd-jni/issues/181#issuecomment-885138495
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Pass the CIs.
Closes#33483 from dongjoon-hyun/SPARK-36262.
Authored-by: Dongjoon Hyun <dongjoon@apache.org>
Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
(cherry picked from commit a1a197403b)
Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
### What changes were proposed in this pull request?
As we decided to release TimestampNTZ type in Spark 3.3, we should update the versions of TimestampNTZ related changes as 3.3.0.
### Why are the changes needed?
Correct the versions in documentation/code comment.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existing UT
Closes#33480 from gengliangwang/updateVersion3.2.
Authored-by: Gengliang Wang <gengliang@apache.org>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
This PR extends the way to represent `delayThreshold` with ANSI interval literals for watermark.
### Why are the changes needed?
A `delayThreshold` is semantically an interval value so it's should be represented as ANSI interval literals as well as the conventional `1 second` form.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
New tests.
Closes#33456 from sarutak/delayThreshold-interval.
Authored-by: Kousuke Saruta <sarutak@oss.nttdata.com>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
(cherry picked from commit 07fa38e2c1)
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Bugfix: link to correction location of Pyspark Dataframe documentation
### Why are the changes needed?
Current website returns "Not found"
### Does this PR introduce _any_ user-facing change?
Website fix
### How was this patch tested?
Documentation change
Closes#33420 from dominikgehl/feature/SPARK-36209.
Authored-by: Dominik Gehl <dog@open.ch>
Signed-off-by: Sean Owen <srowen@gmail.com>
(cherry picked from commit 3a1db2ddd4)
Signed-off-by: Sean Owen <srowen@gmail.com>
### What changes were proposed in this pull request?
This PR proposes removing some APIs from pandas-on-Spark documentation.
Because they can be easily workaround via Spark DataFrame or Column functions, so they might be removed In the future.
### Why are the changes needed?
Because we don't want to expose some functions as a public API.
### Does this PR introduce _any_ user-facing change?
The APIs such as `(Series|Index).spark.data_type`, `(Series|Index).spark.nullable`, `DataFrame.spark.schema`, `DataFrame.spark.print_schema`, `DataFrame.pandas_on_spark.attach_id_column`, `DataFrame.spark.checkpoint`, `DataFrame.spark.localcheckpoint` and `DataFrame.spark.explain` is removed in the documentation.
### How was this patch tested?
Manually build the documents.
Closes#33458 from itholic/SPARK-36239.
Authored-by: itholic <haejoon.lee@databricks.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit 86471ad668)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR follows up #33379 to fix build error in Sphinx
### Why are the changes needed?
The Sphinx build is failed with missing newline in docstring
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Manually test the Sphinx build
Closes#33479 from itholic/SPARK-35810-followup.
Authored-by: itholic <haejoon.lee@databricks.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit d1a037a27c)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR proposes adding an argument `index_col` for `ps.sql` function, to preserve the index when users want.
NOTE that the `reset_index()` have to be performed before using `ps.sql` with `index_col`.
```python
>>> psdf
A B
a 1 4
b 2 5
c 3 6
>>> psdf_reset_index = psdf.reset_index()
>>> ps.sql("SELECT * from {psdf_reset_index} WHERE A > 1", index_col="index")
A B
index
b 2 5
c 3 6
```
Otherwise, the index is always lost.
```python
>>> ps.sql("SELECT * from {psdf} WHERE A > 1")
A B
0 2 5
1 3 6
```
### Why are the changes needed?
Index is one of the key object for the existing pandas users, so we should provide the way to keep the index after computing the `ps.sql`.
### Does this PR introduce _any_ user-facing change?
Yes, the new argument is added.
### How was this patch tested?
Add a unit test and manually check the build pass.
Closes#33450 from itholic/SPARK-35809.
Authored-by: itholic <haejoon.lee@databricks.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit 6578f0b135)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Add `remove_categories` to `CategoricalAccessor` and `CategoricalIndex`.
### Why are the changes needed?
We should implement `remove_categories` in `CategoricalAccessor` and `CategoricalIndex`.
### Does this PR introduce _any_ user-facing change?
Yes, users will be able to use `remove_categories`.
### How was this patch tested?
Added some tests.
Closes#33474 from ueshin/issues/SPARK-36249/remove_categories.
Authored-by: Takuya UESHIN <ueshin@databricks.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit a3c7ae18e2)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
GHA probably doesn't have the same resources as jenkins so move down from 5 to 3 execs and give a bit more time for them to come up.
### Why are the changes needed?
Test is timing out in GHA
### Does this PR introduce _any_ user-facing change?
No, test only change.
### How was this patch tested?
Run through GHA verify no OOM during WorkerDecommissionExtended
Closes#33467 from holdenk/SPARK-36246-WorkerDecommissionExtendedSuite-flakes-in-GHA.
Lead-authored-by: Holden Karau <holden@pigscanfly.ca>
Co-authored-by: Holden Karau <hkarau@netflix.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit 89a83196ac)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Add `add_categories` to `CategoricalAccessor` and `CategoricalIndex`.
### Why are the changes needed?
We should implement `add_categories` in `CategoricalAccessor` and `CategoricalIndex`.
### Does this PR introduce _any_ user-facing change?
Yes, users will be able to use `add_categories`.
### How was this patch tested?
Added some tests.
Closes#33470 from ueshin/issues/SPARK-36214/add_categories.
Authored-by: Takuya UESHIN <ueshin@databricks.com>
Signed-off-by: Takuya UESHIN <ueshin@databricks.com>
(cherry picked from commit dcc0aaa3ef)
Signed-off-by: Takuya UESHIN <ueshin@databricks.com>
### What changes were proposed in this pull request?
This PR adds the version that added pandas API on Spark in PySpark documentation.
### Why are the changes needed?
To document the version added.
### Does this PR introduce _any_ user-facing change?
No to end user. Spark 3.2 is not released yet.
### How was this patch tested?
Linter and documentation build.
Closes#33473 from HyukjinKwon/SPARK-36253.
Authored-by: Hyukjin Kwon <gurwls223@apache.org>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit f3e29574d9)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR adds optimization for scalar and lateral subqueries with OneRowRelation as leaf nodes. It inlines such subqueries before decorrelation to avoid rewriting them as left outer joins. It also introduces a flag to turn on/off this optimization: `spark.sql.optimizer.optimizeOneRowRelationSubquery` (default: True).
For example:
```sql
select (select c1) from t
```
Analyzed plan:
```
Project [scalar-subquery#17 [c1#18] AS scalarsubquery(c1)#22]
: +- Project [outer(c1#18)]
: +- OneRowRelation
+- LocalRelation [c1#18, c2#19]
```
Optimized plan before this PR:
```
Project [c1#18#25 AS scalarsubquery(c1)#22]
+- Join LeftOuter, (c1#24 <=> c1#18)
:- LocalRelation [c1#18]
+- Aggregate [c1#18], [c1#18 AS c1#18#25, c1#18 AS c1#24]
+- LocalRelation [c1#18]
```
Optimized plan after this PR:
```
LocalRelation [scalarsubquery(c1)#22]
```
### Why are the changes needed?
To optimize query plans.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Added new unit tests.
Closes#33284 from allisonwang-db/spark-36063-optimize-subquery-one-row-relation.
Authored-by: allisonwang-db <allison.wang@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
(cherry picked from commit de8e4be92c)
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR partially backports the fix in the script at https://github.com/apache/spark/pull/33410 to make the branch-3.2 build pass at https://github.com/apache/spark/actions/workflows/build_and_test.yml?query=event%3Aschedule
### Why are the changes needed?
To make the Scala 2.13 periodical job pass
### Does this PR introduce _any_ user-facing change?
No, dev-only.
### How was this patch tested?
It is a logically non-conflicting backport.
Closes#33472 from HyukjinKwon/SPARK-36251.
Authored-by: Hyukjin Kwon <gurwls223@apache.org>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This PR upgrades `zstd-jni` from `1.5.0-2` to `1.5.0-3`.
`1.5.0-3` was released few days ago.
This release resolves an issue about buffer size calculation, which can affect usage in Spark.
https://github.com/luben/zstd-jni/releases/tag/v1.5.0-3
### Why are the changes needed?
It might be a corner case that skipping length is greater than `2^31 - 1` but it's possible to affect Spark.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
CI.
Closes#33464 from sarutak/upgrade-zstd-jni-1.5.0-3.
Authored-by: Kousuke Saruta <sarutak@oss.nttdata.com>
Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
(cherry picked from commit dcb7db5370)
Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
### What changes were proposed in this pull request?
Add categories setter to `CategoricalAccessor` and `CategoricalIndex`.
### Why are the changes needed?
We should implement categories setter in `CategoricalAccessor` and `CategoricalIndex`.
### Does this PR introduce _any_ user-facing change?
Yes, users will be able to use categories setter.
### How was this patch tested?
Added some tests.
Closes#33448 from ueshin/issues/SPARK-36188/categories_setter.
Authored-by: Takuya UESHIN <ueshin@databricks.com>
Signed-off-by: Takuya UESHIN <ueshin@databricks.com>
(cherry picked from commit d506815a92)
Signed-off-by: Takuya UESHIN <ueshin@databricks.com>
### What changes were proposed in this pull request?
This PR changes `BaseScriptTransformationExec` for `SparkScriptTransformationExec` to support ANSI interval types.
### Why are the changes needed?
`SparkScriptTransformationExec` support `CalendarIntervalType` so it's better to support ANSI interval types as well.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
New test.
Authored-by: Kousuke Saruta <sarutakoss.nttdata.com>
Signed-off-by: Max Gekk <max.gekkgmail.com>
(cherry picked from commit f56c7b71ff)
Signed-off-by: Max Gekk <max.gekkgmail.com>
Closes#33463 from MaxGekk/sarutak_script-transformation-interval-3.2.
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?
Remove TimestampNTZ type support in the production code of Spark 3.2.
To archive the goal, this PR adds the check "Utils.isTesting" in the following code branches:
- keyword "timestamp_ntz" and "timestamp_ltz" in parser
- New expressions from https://issues.apache.org/jira/browse/SPARK-35662
- Using java.time.localDateTime as the external type for TimestampNTZType
- `SQLConf.timestampType` which determines the default timestamp type of Spark SQL.
This is to minimize the code difference between the master branch. So that future users won't think TimestampNTZ is already available in Spark 3.2.
The downside is that users can still find TimestampNTZType under package `org.apache.spark.sql.types`. There should be nothing left other than this.
### Why are the changes needed?
As of now, there are some blockers for delivering the TimestampNTZ project in Spark 3.2:
- In the Hive Thrift server, both TimestampType and TimestampNTZType are mapped to the same timestamp type, which can cause confusion for users.
- For the Parquet data source, the new written TimestampNTZType Parquet columns will be read as TimestampType in old Spark releases. Also, we need to decide the merge schema for files mixed with TimestampType and TimestampNTZ type.
- The type coercion rules for TimestampNTZType are incomplete. For example, what should the data type of the in clause "IN(Timestamp'2020-01-01 00:00:00', TimestampNtz'2020-01-01 00:00:00') be.
- It is tricky to support TimestampNTZType in JSON/CSV data readers. We need to avoid regressions as possible as we can.
There are 10 days left for the expected 3.2 RC date. So, I propose to **release the TimestampNTZ type in Spark 3.3 instead of Spark 3.2**. So that we have enough time to make considerate designs for the issues.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Existing Unit tests + manual tests from spark-shell to validate the changes are gone.
New functions
```
spark.sql("select to_timestamp_ntz'2021-01-01 00:00:00'").show()
spark.sql("select to_timestamp_ltz'2021-01-01 00:00:00'").show()
spark.sql("select make_timestamp_ntz(1,1,1,1,1,1)").show()
spark.sql("select make_timestamp_ltz(1,1,1,1,1,1)").show()
spark.sql("select localtimestamp()").show()
```
The SQL configuration `spark.sql.timestampType` should not work in 3.2
```
spark.conf.set("spark.sql.timestampType", "TIMESTAMP_NTZ")
spark.sql("select make_timestamp(1,1,1,1,1,1)").schema
spark.sql("select to_timestamp('2021-01-01 00:00:00')").schema
spark.sql("select timestamp'2021-01-01 00:00:00'").schema
Seq((1, java.sql.Timestamp.valueOf("2021-01-01 00:00:00"))).toDF("i", "ts").write.partitionBy("ts").parquet("/tmp/test")
spark.read.parquet("/tmp/test").schema
```
LocalDateTime is not supported as a built-in external type:
```
Seq(LocalDateTime.now()).toDF()
org.apache.spark.sql.catalyst.expressions.Literal(java.time.LocalDateTime.now())
org.apache.spark.sql.catalyst.expressions.Literal(0L, TimestampNTZType)
```
Closes#33444 from gengliangwang/banNTZ.
Authored-by: Gengliang Wang <gengliang@apache.org>
Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
### What changes were proposed in this pull request?
This fixes a case sensitivity issue for desc table commands with partition specified.
### Why are the changes needed?
bugfix
### Does this PR introduce _any_ user-facing change?
yes, but it's a bugfix
### How was this patch tested?
new tests
#### before
```
+-- !query
+DESC EXTENDED t PARTITION (C='Us', D=1)
+-- !query schema
+struct<>
+-- !query output
+org.apache.spark.sql.AnalysisException
+Partition spec is invalid. The spec (C, D) must match the partition spec (c, d) defined in table '`default`.`t`'
+
```
#### after
https://github.com/apache/spark/pull/33424/files#diff-554189c49950974a948f99fa9b7436f615052511660c6a0ae3062fa8ca0a327cR328Closes#33424 from yaooqinn/SPARK-36213.
Authored-by: Kent Yao <yao@apache.org>
Signed-off-by: Kent Yao <yao@apache.org>
(cherry picked from commit 4cd6cfc773)
Signed-off-by: Kent Yao <yao@apache.org>
### What changes were proposed in this pull request?
For non-datasource Hive tables, e.g. tables written outside of Spark (through Hive or Trino), we have certain optimzations in Spark where we use Spark ORC and Parquet datasources to read these tables ([Ref](fbf53dee37/sql/hive/src/main/scala/org/apache/spark/sql/hive/HiveMetastoreCatalog.scala (L128))) rather than using the Hive serde.
If such a table contains a `path` property, Spark will try to list this path property in addition to the table location when creating an `InMemoryFileIndex`. ([Ref](fbf53dee37/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/DataSource.scala (L575))) This can lead to wrong data if `path` property points to a directory location or an error if `path` is not a location. A concrete example is provided in [SPARK-28266 (comment)](https://issues.apache.org/jira/browse/SPARK-28266?focusedCommentId=17380170&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-17380170).
Since these tables were not written through Spark, Spark should not interpret this `path` property as it can be set by an external system with a different meaning.
### Why are the changes needed?
For better compatibility with Hive tables generated by other platforms (non-Spark)
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Added unit test
Closes#33328 from shardulm94/spark-28266.
Authored-by: Shardul Mahadik <smahadik@linkedin.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
(cherry picked from commit 685c3fd05b)
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Sometimes, AQE skew join optimization can fail with NPE. This is because AQE tries to get the shuffle block sizes, but some map outputs are missing due to the executor lost or something.
This PR fixes this bug by skipping skew join handling if some map outputs are missing in the `MapOutputTracker`.
### Why are the changes needed?
bug fix
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
a new UT
Closes#33445 from cloud-fan/bug.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
(cherry picked from commit 9c8a3d3975)
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Spark 3.2.0 will use parquet-mr.1.12.0 version (or higher), that contains the column encryption feature which can be called from Spark SQL. The aim of this PR is to document the use of Parquet encryption in Spark.
### Why are the changes needed?
- To provide information on how to use Parquet column encryption
### Does this PR introduce _any_ user-facing change?
Yes, documents a new feature.
### How was this patch tested?
bundle exec jekyll build
Closes#32895 from ggershinsky/parquet-encryption-doc.
Authored-by: Gidon Gershinsky <ggershinsky@apple.com>
Signed-off-by: Sean Owen <srowen@gmail.com>
### What changes were proposed in this pull request?
Update trasform's doc to latest code.
### Why are the changes needed?
keep consistence
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
No
Closes#33362 from AngersZhuuuu/SPARK-36153.
Lead-authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Co-authored-by: AngersZhuuuu <angers.zhu@gmail.com>
Signed-off-by: Sean Owen <srowen@gmail.com>
### What changes were proposed in this pull request?
This is a followup of https://github.com/apache/spark/pull/33222 .
https://github.com/apache/spark/pull/33222 made a mistake that, `RemoveRedundantProjects` may lose the `LOGICAL_PLAN_TAG` tag, even though the logical plan link is retained. This was actually caught by the test `LogicalPlanTagInSparkPlanSuite`, but was not being taken care of.
There is no problem so far, but losing information can always lead to potential bugs.
### Why are the changes needed?
fix a mistake
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
existing test
Closes#33442 from cloud-fan/minor.
Authored-by: Wenchen Fan <wenchen@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
(cherry picked from commit 94aece4325)
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Adding support for accepting an initial state with flatMapGroupsWithState in batch mode.
### Why are the changes needed?
SPARK-35897 added support for accepting an initial state for streaming queries using flatMapGroupsWithState. the code flow is separate for batch and streaming and required a different PR.
### Does this PR introduce _any_ user-facing change?
Yes as discussed above flatMapGroupsWithState in batch mode can accept an initialState, previously this would throw an UnsupportedOperationException
### How was this patch tested?
Added relevant unit tests in FlatMapGroupsWithStateSuite and modified the tests `JavaDatasetSuite`
Closes#33336 from rahulsmahadev/flatMapGroupsWithStateBatch.
Authored-by: Rahul Mahadev <rahul.mahadev@databricks.com>
Signed-off-by: Tathagata Das <tathagata.das1565@gmail.com>
(cherry picked from commit efcce23b91)
Signed-off-by: Tathagata Das <tathagata.das1565@gmail.com>
### What changes were proposed in this pull request?
Removes `FileFormatDataWriterMetricSuite` which duplicated.
### Why are the changes needed?
`FileFormatDataWriterMetricSuite` should be renamed to `InMemoryTableMetricSuite`. But it was wrongly copied.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Existing tests.
Closes#33454 from viirya/SPARK-36030-followup-3.2.
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?
This PR avoid using procedure syntax deprecated in Scala 2.13.
https://github.com/apache/spark/runs/3120481756?check_suite_focus=true
```
[error] /home/runner/work/spark/spark/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/FileFormatDataWriterMetricSuite.scala:44:90: procedure syntax is deprecated: instead, add `: Unit =` to explicitly declare `testMetricOnDSv2`'s return type
[error] private def testMetricOnDSv2(func: String => Unit, checker: Map[Long, String] => Unit) {
[error] ^
[error] /home/runner/work/spark/spark/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/InMemoryTableMetricSuite.scala:44:90: procedure syntax is deprecated: instead, add `: Unit =` to explicitly declare `testMetricOnDSv2`'s return type
[error] private def testMetricOnDSv2(func: String => Unit, checker: Map[Long, String] => Unit) {
[error] ^
[warn] 100 warnings found
[error] two errors found
[error] (sql / Test / compileIncremental) Compilation failed
[error] Total time: 579 s (09:39), completed Jul 21, 2021 4:14:26 AM
```
### Why are the changes needed?
To make the build compatible with Scala 2.13 in Spark.
### Does this PR introduce _any_ user-facing change?
No, dev-only.
### How was this patch tested?
Manually tested:
```bash
./dev/change-scala-version.sh 2.13
./build/mvn -DskipTests -Phive-2.3 -Phive clean package -Pscala-2.13
```
Closes#33452 from HyukjinKwon/SPARK-36030.
Authored-by: Hyukjin Kwon <gurwls223@apache.org>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit 99006e515b)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
We add the interface for DS v2 metrics in SPARK-34366. It is only added for reading path, though. This patch extends the metrics interface to writing path.
### Why are the changes needed?
Complete DS v2 metrics interface support in writing path.
### Does this PR introduce _any_ user-facing change?
No. For developer, yes, as this adds metrics support at DS v2 writing path.
### How was this patch tested?
Added test.
Closes#33239 from viirya/v2-write-metrics.
Authored-by: Liang-Chi Hsieh <viirya@gmail.com>
Signed-off-by: Liang-Chi Hsieh <viirya@gmail.com>
(cherry picked from commit 2653201b0a)
Signed-off-by: Liang-Chi Hsieh <viirya@gmail.com>
### What changes were proposed in this pull request?
1. add "closeStreams" to FileAppender and RollingFileAppender
2. set "closeStreams" to "true" in ExecutorRunner
### Why are the changes needed?
The executor will hang when due disk full or other exceptions which happened in writting to outputStream: the root cause is the "inputStream" is not closed after the error happens:
1. ExecutorRunner creates two files appenders for pipe: one for stdout, one for stderr
2. FileAppender.appendStreamToFile exits the loop when writing to outputStream
3. FileAppender closes the outputStream, but left the inputStream which refers the pipe's stdout and stderr opened
4. The executor will hang when printing the log message if the pipe is full (no one consume the outputs)
5. From the driver side, you can see the task can't be completed for ever
With this fix, the step 4 will throw an exception, the driver can catch up the exception and reschedule the failed task to other executors.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Add new tests for the "closeStreams" in FileAppenderSuite
Closes#33263 from jhu-chang/SPARK-35027.
Authored-by: Jie <gt.hu.chang@gmail.com>
Signed-off-by: Sean Owen <srowen@gmail.com>
(cherry picked from commit 1a8c6755a1)
Signed-off-by: Sean Owen <srowen@gmail.com>
### What changes were proposed in this pull request?
This PR documents a new feature "native support of session window" into Structured Streaming guide doc.
Screenshots are following:
### Why are the changes needed?
This change is needed to explain a new feature to the end users.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Documentation changes.
Closes#33433 from HeartSaVioR/SPARK-36172.
Authored-by: Jungtaek Lim <kabhwan.opensource@gmail.com>
Signed-off-by: Jungtaek Lim <kabhwan.opensource@gmail.com>
(cherry picked from commit 0eb31a06d6)
Signed-off-by: Jungtaek Lim <kabhwan.opensource@gmail.com>
### What changes were proposed in this pull request?
Add `as_ordered`/`as_unordered` to `CategoricalAccessor` and `CategoricalIndex`.
### Why are the changes needed?
We should implement `as_ordered`/`as_unordered` in `CategoricalAccessor` and `CategoricalIndex` yet.
### Does this PR introduce _any_ user-facing change?
Yes, users will be able to use `as_ordered`/`as_unordered`.
### How was this patch tested?
Added some tests.
Closes#33400 from ueshin/issues/SPARK-36186/as_ordered_unordered.
Authored-by: Takuya UESHIN <ueshin@databricks.com>
Signed-off-by: Takuya UESHIN <ueshin@databricks.com>
(cherry picked from commit 376fadc89c)
Signed-off-by: Takuya UESHIN <ueshin@databricks.com>
### What changes were proposed in this pull request?
The current implement of `Sequence` expression not support step by days for dates.
```
spark-sql> select sequence(date'2021-07-01', date'2021-07-10', interval '3' day);
Error in query: cannot resolve 'sequence(DATE '2021-07-01', DATE '2021-07-10', INTERVAL '3' DAY)' due to data type mismatch:
sequence uses the wrong parameter type. The parameter type must conform to:
1. The start and stop expressions must resolve to the same type.
2. If start and stop expressions resolve to the 'date' or 'timestamp' type
then the step expression must resolve to the 'interval' or
'interval year to month' or 'interval day to second' type,
otherwise to the same type as the start and stop expressions.
; line 1 pos 7;
'Project [unresolvedalias(sequence(2021-07-01, 2021-07-10, Some(INTERVAL '3' DAY), Some(Europe/Moscow)), None)]
+- OneRowRelation
```
### Why are the changes needed?
`DayTimeInterval` has day granularity should as step for dates.
### Does this PR introduce _any_ user-facing change?
'Yes'.
Sequence expression will supports step by `DayTimeInterval` has day granularity for dates.
### How was this patch tested?
New tests.
Closes#33439 from beliefer/SPARK-36222.
Authored-by: gengjiaan <gengjiaan@360.cn>
Signed-off-by: Max Gekk <max.gekk@gmail.com>
(cherry picked from commit c0d84e6cf1)
Signed-off-by: Max Gekk <max.gekk@gmail.com>
### What changes were proposed in this pull request?
Preserve the insertion order of columns in Dataset.withColumns
### Why are the changes needed?
It is the expected behavior. We preserve insertion order in all other places.
### Does this PR introduce _any_ user-facing change?
No. Currently Dataset.withColumns is not actually used anywhere to insert more than one column. This change is to make sure it behaves as expected when it is used for that purpose in future.
### How was this patch tested?
Added test in DatasetSuite
Closes#33423 from koertkuipers/feat-withcolumns-preserve-order.
Authored-by: Koert Kuipers <koert@tresata.com>
Signed-off-by: Liang-Chi Hsieh <viirya@gmail.com>
(cherry picked from commit bf680bf25a)
Signed-off-by: Liang-Chi Hsieh <viirya@gmail.com>
### What changes were proposed in this pull request?
Forces the selectivity estimate for null-based filters to be in the range `[0,1]`.
### Why are the changes needed?
I noticed in a few TPC-DS query tests that the column statistic null count can be higher than the table statistic row count. In the current implementation, the selectivity estimate for `IsNotNull` is negative.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Unit test
Closes#33286 from karenfeng/bound-selectivity-est.
Authored-by: Karen Feng <karen.feng@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
(cherry picked from commit ddc61e62b9)
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR follows https://github.com/apache/spark/pull/33299 and implement `prettyName` for `MakeTimestampNTZ` and `MakeTimestampLTZ` based on the discussion show below
https://github.com/apache/spark/pull/33299/files#r668423810
### Why are the changes needed?
This PR fix the incorrect alias usecase.
### Does this PR introduce _any_ user-facing change?
'No'.
Modifications are transparent to users.
### How was this patch tested?
Jenkins test.
Closes#33430 from beliefer/SPARK-36046-followup.
Authored-by: gengjiaan <gengjiaan@360.cn>
Signed-off-by: Gengliang Wang <gengliang@apache.org>
(cherry picked from commit 033a5731b4)
Signed-off-by: Gengliang Wang <gengliang@apache.org>
### What changes were proposed in this pull request?
When inner field have wrong schema filed name should check field name too.
### Why are the changes needed?
Early check early faield
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Added UT
Closes#33409 from AngersZhuuuu/SPARK-36201.
Authored-by: Angerszhuuuu <angers.zhu@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
(cherry picked from commit 251885772d)
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
* Add non-empty partition check in `CustomShuffleReaderExec`
* Make sure `OptimizeLocalShuffleReader` doesn't return empty partition
### Why are the changes needed?
Since SPARK-32083, AQE coalesce always return at least one partition, it should be robust to add non-empty check in `CustomShuffleReaderExec`.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
not need
Closes#33431 from ulysses-you/non-empty-partition.
Authored-by: ulysses-you <ulyssesyou18@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
(cherry picked from commit b70c25881c)
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR documents about unicode literals added in SPARK-34051 (#31096) and a past PR in `sql-ref-literals.md`.
### Why are the changes needed?
Notice users about the literals.
### Does this PR introduce _any_ user-facing change?
Yes, but just add a sentence.
### How was this patch tested?
Built the document and confirmed the result.
```
SKIP_API=1 bundle exec jekyll build
```
Closes#33434 from sarutak/unicode-literal-doc.
Authored-by: Kousuke Saruta <sarutak@oss.nttdata.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
(cherry picked from commit ba1294ea5a)
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This is one of the patches for SPIP SPARK-30602 which is needed for push-based shuffle.
### Summary of the change:
When Executor registers with Shuffle Service, it will encode the merged shuffle dir created and also the application attemptId into the ShuffleManagerMeta into Json. Then in Shuffle Service, it will decode the Json string and get the correct merged shuffle dir and also the attemptId. If the registration comes from a newer attempt, the merged shuffle information will be updated to store the information from the newer attempt.
This PR also refactored the management of the merged shuffle information to avoid concurrency issues.
### 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.
Closes#33078 from zhouyejoe/SPARK-35546.
Authored-by: Ye Zhou <yezhou@linkedin.com>
Signed-off-by: Mridul Muralidharan <mridul<at>gmail.com>
(cherry picked from commit c77acf0bbc)
Signed-off-by: Mridul Muralidharan <mridulatgmail.com>
### What changes were proposed in this pull request?
Test is flaky (https://github.com/apache/spark/runs/3109815586):
```
Traceback (most recent call last):
File "/__w/spark/spark/python/pyspark/mllib/tests/test_streaming_algorithms.py", line 391, in test_parameter_convergence
eventually(condition, catch_assertions=True)
File "/__w/spark/spark/python/pyspark/testing/utils.py", line 91, in eventually
raise lastValue
File "/__w/spark/spark/python/pyspark/testing/utils.py", line 82, in eventually
lastValue = condition()
File "/__w/spark/spark/python/pyspark/mllib/tests/test_streaming_algorithms.py", line 387, in condition
self.assertEqual(len(model_weights), len(batches))
AssertionError: 9 != 10
```
Should probably increase timeout
### Why are the changes needed?
To avoid flakiness in the test.
### Does this PR introduce _any_ user-facing change?
Nope, dev-only.
### How was this patch tested?
CI should test it out.
Closes#33427 from HyukjinKwon/SPARK-36216.
Authored-by: Hyukjin Kwon <gurwls223@apache.org>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit d6b974f8ce)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Support TimestampNTZType in SparkGetColumnsOperation
### Why are the changes needed?
TimestampNTZType coverage
### Does this PR introduce _any_ user-facing change?
yes, jdbc end-users will be aware of TimestampNTZType
### How was this patch tested?
add new test
Closes#33393 from yaooqinn/SPARK-36179.
Authored-by: Kent Yao <yao@apache.org>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
(cherry picked from commit 0c76fb9c01)
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
This is a backport of #33377.
Revisit and manage `InternalField` in more places.
### Why are the changes needed?
There are other places we can manage `InternalField`, and we can keep extension dtypes or `CategoricalDtype`.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Added some tests.
Closes#33384 from ueshin/issues/SPARK-36167/3.2/internal_field.
Authored-by: Takuya UESHIN <ueshin@databricks.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Support comparison between a Categorical and a scalar.
There are 3 main changes:
- Modify `==` and `!=` from comparing **codes** of the Categorical to the scalar to comparing **actual values** of the Categorical to the scalar.
- Support `<`, `<=`, `>`, `>=` between a Categorical and a scalar.
- TypeError message fix.
### Why are the changes needed?
pandas supports comparison between a Categorical and a scalar, we should follow pandas' behaviors.
### Does this PR introduce _any_ user-facing change?
Yes.
Before:
```py
>>> import pyspark.pandas as ps
>>> import pandas as pd
>>> from pandas.api.types import CategoricalDtype
>>> pser = pd.Series(pd.Categorical([1, 2, 3], categories=[3, 2, 1], ordered=True))
>>> psser = ps.from_pandas(pser)
>>> psser == 2
0 True
1 False
2 False
dtype: bool
>>> psser <= 1
Traceback (most recent call last):
...
NotImplementedError: <= can not be applied to categoricals.
```
After:
```py
>>> import pyspark.pandas as ps
>>> import pandas as pd
>>> from pandas.api.types import CategoricalDtype
>>> pser = pd.Series(pd.Categorical([1, 2, 3], categories=[3, 2, 1], ordered=True))
>>> psser = ps.from_pandas(pser)
>>> psser == 2
0 False
1 True
2 False
dtype: bool
>>> psser <= 1
0 True
1 True
2 True
dtype: bool
```
### How was this patch tested?
Unit tests.
Closes#33373 from xinrong-databricks/categorical_eq.
Authored-by: Xinrong Meng <xinrong.meng@databricks.com>
Signed-off-by: Takuya UESHIN <ueshin@databricks.com>
(cherry picked from commit 8dd43351d5)
Signed-off-by: Takuya UESHIN <ueshin@databricks.com>
