### What changes were proposed in this pull request?
Follow up on [SPARK-30428](https://github.com/apache/spark/pull/27112) which added support for partition pruning in File source V2.
This PR implements the necessary changes in order to pass the `dataFilters` to the `listFiles`. This enables having `FileIndex` implementations which use the `dataFilters` for further pruning the file listing (see the discussion [here](https://github.com/apache/spark/pull/27112#discussion_r364757217)).
### Why are the changes needed?
Datasources such as `csv` and `json` do not implement the `SupportsPushDownFilters` trait. In order to support data skipping uniformly for all file based data sources, one can override the `listFiles` method in a `FileIndex` implementation, which consults external metadata and prunes the list of files.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Modifying the unit tests for v2 file sources to verify the `dataFilters` are passed
Closes#27157 from guykhazma/PushdataFiltersInFileListing.
Authored-by: Guy Khazma <guykhag@gmail.com>
Signed-off-by: Gengliang Wang <gengliang.wang@databricks.com>
### What changes were proposed in this pull request?
In the PR, I propose to make `JsonSuite` and `CSVSuite` abstract classes, and add sub-classes that check JSON/CSV datasource v1 and v2.
### Why are the changes needed?
To improve test coverage and test JSON/CSV v1 which is still supported, and can be enabled by users.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
By running new test suites `JsonV1Suite` and `CSVv1Suite`.
Closes#27294 from MaxGekk/csv-json-v1-test-suites.
Authored-by: Maxim Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
After this commit d67b98ea01, we are able to create table or alter table with interval column types if the external catalog accepts which is varying the interval type's purpose for internal usage. With d67b98ea01 's original purpose it should only work from cast logic.
Instead of adding type checker for the interval type from commands to commands to work among catalogs, It much simpler to treat interval as an invalid data type but can be identified by cast only.
### Why are the changes needed?
enhance interval internal usage purpose.
### Does this PR introduce any user-facing change?
NO,
Additionally, this PR restores user behavior when using interval type to create/alter table schema, e.g. for hive catalog
for 2.4,
```java
Caused by: org.apache.spark.sql.catalyst.parser.ParseException:
DataType calendarinterval is not supported.(line 1, pos 0)
```
for master after d67b98ea01
```java
Caused by: org.apache.hadoop.hive.ql.metadata.HiveException: java.lang.IllegalArgumentException: Error: type expected at the position 0 of 'interval' but 'interval' is found.
at org.apache.hadoop.hive.ql.metadata.Hive.createTable(Hive.java:862)
```
now with this pr, we restore the type checker in spark side.
### How was this patch tested?
add more ut
Closes#27277 from yaooqinn/SPARK-30568.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Add `owner` property to v2 table, it is reversed by `TableCatalog`, indicates the table's owner.
### Why are the changes needed?
enhance ownership management of catalog API
### Does this PR introduce any user-facing change?
yes, add 1 reserved property - `owner` , and it is not allowed to use in OPTIONS/TBLPROPERTIES anymore, only if legacy on
### How was this patch tested?
add uts
Closes#27249 from yaooqinn/SPARK-30019.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR proposes to clean up `UnivocityParser`.
### Why are the changes needed?
It will slightly improve the performance since we don't do unnecessary computation for Array concatenations/creation.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Manually ran the existing tests.
Closes#27287 from HyukjinKwon/SPARK-30530-followup.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
In the PR, I propose to output additional msg from the tests where a log appender is added. The message is printed as a part of `IllegalStateException` in the case of reaching the limit of maximum number of logged events.
### Why are the changes needed?
If a log appender is not removed from the log4j appenders list. the caller message could help to investigate the problem and find the test which doesn't remove the log appender.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
By running the modified test suites `AvroSuite`, `CSVSuite`, `ResolveHintsSuite` and etc.
Closes#27296 from MaxGekk/assign-name-to-log-appender.
Authored-by: Maxim Gekk <max.gekk@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Explicitly set conf to let orc use DSv2 in `OrcFilterSuite` in both v1.2 and v2.3.
### Why are the changes needed?
Tests should not rely on default conf when they're going to test something intentionally, which can be fail when conf changes.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Pass Jenkins.
Closes#27285 from Ngone51/fix-orcfilter-test.
Authored-by: yi.wu <yi.wu@databricks.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Use the new framework to resolve the ALTER TABLE commands.
This PR also refactors ALTER TABLE logical plans such that they extend a base class `AlterTable`. Each plan now implements `def changes: Seq[TableChange]` for any table change operations.
Additionally, `UnresolvedV2Relation` and its usage is completely removed.
### Why are the changes needed?
This is a part of effort to make the relation lookup behavior consistent: [SPARK-29900](https://issues.apache.org/jira/browse/SPARK-29900).
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Updated existing tests
Closes#27243 from imback82/v2commands_newframework.
Authored-by: Terry Kim <yuminkim@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
In the PR, I propose move out creation of `AvroOption` from `AvroPartitionReaderFactory.buildReader`, and create it earlier in `AvroScan.createReaderFactory`.
### Why are the changes needed?
- To avoid building `AvroOptions` from a map of Avro options and Hadoop conf per each partition.
- If an instance of `AvroOptions` is built only once at the driver side, we could output warnings while parsing Avro options and don't worry about noisiness of the warnings.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
By `AvroSuite`
Closes#27272 from MaxGekk/avro-options-once-for-read.
Authored-by: Maxim Gekk <max.gekk@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Changed signature of `rawParser` passed to `FailureSafeParser`. I propose to change return type from `Seq` to `Iterable`. I took `Iterable` to easier port the changes on Scala collections 2.13. Also, I replaced `Seq` by `Option` in CSV datasource - `UnivocityParser`, and in JSON parser exception one place in the case when specified schema is `StructType`, and JSON input is an array.
### Why are the changes needed?
`Seq` is unnecessary requirement for return type from rawParser which may not have multiple rows per input like CSV datasource.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
By existing test suites `JsonSuite`, `UnivocityParserSuite`, `JsonFunctionsSuite`, `JsonExpressionsSuite`, `CsvSuite`, and `CsvFunctionsSuite`.
Closes#27264 from MaxGekk/failuresafe-parser-seq.
Authored-by: Maxim Gekk <max.gekk@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
`CalendarInterval` is maintained as a private class but might be used in a public way by users
e.g.
```scala
scala> spark.udf.register("getIntervalMonth", (_:org.apache.spark.unsafe.types.CalendarInterval).months)
scala> sql("select interval 2 month 1 day a").selectExpr("getIntervalMonth(a)").show
+-------------------+
|getIntervalMonth(a)|
+-------------------+
| 2|
+-------------------+
```
And it exists since 1.5.0, now we go to the 3.x era，may be it's time to make it public
### Why are the changes needed?
make the interval more future-proofing
### Does this PR introduce any user-facing change?
doc change
### How was this patch tested?
add ut.
Closes#27258 from yaooqinn/SPARK-30547.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
This PR implements a tiny performance optimization for a `GenericArrayData` constructor, avoiding an unnecessary roundtrip through `WrappedArray` when the provided value is already an array of objects.
It also fixes a related performance problem in `ParquetRowConverter`.
### Why are the changes needed?
`GenericArrayData` has a `this(seqOrArray: Any)` constructor, which was originally added in #13138 for use in `RowEncoder` (where we may not know concrete types until runtime) but is also called (perhaps unintentionally) in a few other code paths.
In this constructor's existing implementation, a call to `new WrappedArray(Array[Object](""))` is dispatched to the `this(seqOrArray: Any)` constructor, where we then call `this(array.toSeq)`: this wraps the provided array into a `WrappedArray`, which is subsequently unwrapped in a `this(seq.toArray)` call. For an interactive example, see https://scastie.scala-lang.org/7jOHydbNTaGSU677FWA8nA
This PR changes the `this(seqOrArray: Any)` constructor so that it calls the primary `this(array: Array[Any])` constructor, allowing us to save a `.toSeq.toArray` call; this comes at the cost of one additional `case` in the `match` statement (but I believe this has a negligible performance impact relative to the other savings).
As code cleanup, I also reverted the JVM 1.7 workaround from #14271.
I also fixed a related performance problem in `ParquetRowConverter`: previously, this code called `ArrayBasedMapData.apply` which, in turn, called the `this(Any)` constructor for `GenericArrayData`: this PR's micro-benchmarks show that this is _significantly_ slower than calling the `this(Array[Any])` constructor (and I also observed time spent here during other Parquet scan benchmarking work). To fix this performance problem, I replaced the call to the `ArrayBasedMapData.apply` method with direct calls to the `ArrayBasedMapData` and `GenericArrayData` constructors.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
I tested this by running code in a debugger and by running microbenchmarks (which I've added to a new `GenericArrayDataBenchmark` in this PR):
- With JDK8 benchmarks: this PR's changes more than double the performance of calls to the `this(Any)` constructor. Even after improvements, however, calls to the `this(Array[Any])` constructor are still ~60x faster than calls to `this(Any)` when passing a non-primitive array (thereby motivating this patch's other change in `ParquetRowConverter`).
- With JDK11 benchmarks: the changes more-or-less completely eliminate the performance penalty associated with the `this(Any)` constructor.
Closes#27088 from JoshRosen/joshrosen/GenericArrayData-optimization.
Authored-by: Josh Rosen <rosenville@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This pr intends to upgrade lz4-java from 1.7.0 to 1.7.1.
### Why are the changes needed?
This release includes a bug fix for older macOS. You can see the link below for the changes;
https://github.com/lz4/lz4-java/blob/master/CHANGES.md#171
### Does this PR introduce any user-facing change?
### How was this patch tested?
Existing tests.
Closes#27271 from maropu/SPARK-30486.
Authored-by: Takeshi Yamamuro <yamamuro@apache.org>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Update Twitter Chill to 0.9.5.
### Why are the changes needed?
Primarily, Scala 2.13 support for later.
Other changes from 0.9.3 are apparently just minor fixes and improvements:
https://github.com/twitter/chill/releases
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Existing tests
Closes#27227 from srowen/SPARK-29290.
Authored-by: Sean Owen <srowen@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This PR aims to add a fallback Maven repository when a mirror to `central` fail.
### Why are the changes needed?
We use `Google Maven Central` in GitHub Action as a mirror of `central`.
However, `Google Maven Central` sometimes doesn't have newly published artifacts
and there is no guarantee when we get the newly published artifacts.
By duplicating `Maven Central` with a new ID, we can add a fallback Maven repository
which is not mirrored by `Google Maven Central`.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Manually testing with the new `Twitter` chill artifacts by switching `chill.version` from `0.9.3` to `0.9.5`.
```
$ rm -rf ~/.m2/repository/com/twitter/chill*
$ mvn compile | grep chill
Downloading from google-maven-central: https://maven-central.storage-download.googleapis.com/repos/central/data/com/twitter/chill_2.12/0.9.5/chill_2.12-0.9.5.pom
Downloading from central_without_mirror: https://repo.maven.apache.org/maven2/com/twitter/chill_2.12/0.9.5/chill_2.12-0.9.5.pom
Downloaded from central_without_mirror: https://repo.maven.apache.org/maven2/com/twitter/chill_2.12/0.9.5/chill_2.12-0.9.5.pom (2.8 kB at 11 kB/s)
```
Closes#27281 from dongjoon-hyun/SPARK-30572.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Renamed an identifier `iterator` to `iter` to avoid compile error with Scala 2.13.
### Why are the changes needed?
As of Scala 2.13, scala.collection.Iterator has "iterator" method so if an inner class of Iterator means to refer an outer identifier named "iterator", it does not work as we think.
I listed source files that can be affected by that change by `find . -name "*.scala" -exec grep -El "new .*Iterator\[.* +{" {} \;`
As far as I confirmed util.Utils` is affected.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Existing tests.
Closes#27275 from sarutak/fix-iterator-for-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 adds a migration guide for `SHOW TBLPROPERTIES` for Apache Spark 3.0.0.
### Why are the changes needed?
The behavior of `SHOW TBLPROPERTIES` changed when the table does not exist. The migration guide reflects this user facing change.
### Does this PR introduce any user-facing change?
Yes. This is a documentation change.
### How was this patch tested?
No tests were added because this is a doc change.
Closes#27276 from imback82/SPARK-30282-followup.
Authored-by: Terry Kim <yuminkim@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
make KUBERNETES_MASTER_INTERNAL_URL configurable
### Why are the changes needed?
we do not always use the default port number 443 to access our kube-apiserver, and even in some mulit-tenant cluster, people do not use the service `kubernetes.default.svc` to access the kube-apiserver, so make the internal master configurable is necessary。
### Does this PR introduce any user-facing change?
user can configure the internal master url by
```
--conf spark.kubernetes.internal.master=https://kubernetes.default.svc:6443
```
### How was this patch tested?
run in multi-cluster that do not use the https://kubernetes.default.svc to access the kube-apiserver
Closes#27029 from wackxu/internalmaster.
Authored-by: xushiwei 00425595 <xushiwei5@huawei.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
In the PR, I propose to fix the bug reported in SPARK-30530. CSV datasource returns invalid records in the case when `parsedSchema` is shorter than number of tokens returned by UniVocity parser. In the case `UnivocityParser.convert()` always throws `BadRecordException` independently from the result of applying filters.
For the described case, I propose to save the exception in `badRecordException` and continue value conversion according to `parsedSchema`. If a bad record doesn't pass filters, `convert()` returns empty Seq otherwise throws `badRecordException`.
### Why are the changes needed?
It fixes the bug reported in the JIRA ticket.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Added new test from the JIRA ticket.
Closes#27239 from MaxGekk/spark-30530-csv-filter-is-null.
Authored-by: Maxim Gekk <max.gekk@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
As we are not going to follow ANSI to implement year-month and day-time interval types, it is weird to compare the year-month part to the day-time part for our current implementation of interval type now.
Additionally, the current ordering logic comes from PostgreSQL where the implementation of the interval is messy. And we are not aiming PostgreSQL compliance at all.
THIS PR will revert https://github.com/apache/spark/pull/26681 and https://github.com/apache/spark/pull/26337
### Why are the changes needed?
make interval type more future-proofing
### Does this PR introduce any user-facing change?
there are new in 3.0, so no
### How was this patch tested?
existing uts shall work
Closes#27262 from yaooqinn/SPARK-30551.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Resolve the remaining comments in [PR#27226](https://github.com/apache/spark/pull/27226).
### Why are the changes needed?
Resolve the comments.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Existing unit tests.
Closes#27253 from JkSelf/followup-skewjoinoptimization2.
Authored-by: jiake <ke.a.jia@intel.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Pick up the HTTP fix from https://issues.apache.org/jira/browse/HIVE-22708
### Why are the changes needed?
This is a small but important fix to digest handling we should pick up from Hive.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Existing tests
Closes#27273 from srowen/Hive22708.
Authored-by: Sean Owen <srowen@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Removes suggestion to use -T for parallel Maven build.
### Why are the changes needed?
Parallel builds don't necessarily work in the build right now.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
N/A
Closes#27274 from srowen/ParallelBuild.
Authored-by: Sean Owen <srowen@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
https://github.com/apache/spark/pull/26809 added `Dataset.tail` API. It should be good to have it in PySpark API as well.
### Why are the changes needed?
To support consistent APIs.
### Does this PR introduce any user-facing change?
No. It adds a new API.
### How was this patch tested?
Manually tested and doctest was added.
Closes#27251 from HyukjinKwon/SPARK-30539.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
Upgrade the version of Genjavadoc from 0.14 to 0.15.
### Why are the changes needed?
To enable to build for Scala 2.13.1.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
I confirmed there is no dependency error related to genjavadoc by manual build.
Also, I generated javadoc by `LANG=C build/sbt -Pkinesis-asl -Pyarn -Pkubernetes -Phive-thriftserver unidoc` for both code with/without this change and did `diff -r` target/javadoc.
Closes#27255 from sarutak/upgrade-genjavadoc.
Authored-by: Kousuke Saruta <sarutak@oss.nttdata.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This PR adds:
- `pyspark.ml.regression.JavaRegressor`
- `pyspark.ml.regression.JavaRegressionModel`
classes and replaces `JavaPredictor` and `JavaPredictionModel` in
- `LinearRegression` / `LinearRegressionModel`
- `DecisionTreeRegressor` / `DecisionTreeRegressionModel` (just addition as `JavaPredictionModel` hasn't been used)
- `RandomForestRegressor` / `RandomForestRegressionModel` (just addition as `JavaPredictionModel` hasn't been used)
- `GBTRegressor` / `GBTRegressionModel` (just addition as `JavaPredictionModel` hasn't been used)
- `AFTSurvivalRegression` / `AFTSurvivalRegressionModel`
- `GeneralizedLinearRegression` / `GeneralizedLinearRegressionModel`
- `FMRegressor` / `FMRegressionModel`
### Why are the changes needed?
- Internal PySpark consistency.
- Feature parity with Scala.
- Intermediate step towards implementing [SPARK-29212](https://issues.apache.org/jira/browse/SPARK-29212)
### Does this PR introduce any user-facing change?
It adds new base classes, so it will affect `mro`. Otherwise interfaces should stay intact.
### How was this patch tested?
Existing tests.
Closes#27241 from zero323/SPARK-30533.
Authored-by: zero323 <mszymkiewicz@gmail.com>
Signed-off-by: Sean Owen <srowen@gmail.com>
### What changes were proposed in this pull request?
This PR adds a migration guide for MsSQLServer JDBC dialect for Apache Spark 2.4.4 and 2.4.5.
### Why are the changes needed?
Apache Spark 2.4.4 updates the type mapping correctly according to MS SQL Server, but missed to mention that in the migration guide. In addition, 2.4.4 adds a configuration for the legacy behavior.
### Does this PR introduce any user-facing change?
Yes. This is a documentation change.
### How was this patch tested?
Manually generate and see the doc.
Closes#27270 from dongjoon-hyun/SPARK-28152-DOC.
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 aims to add these test cases for resolution of ORC table location reported by [SPARK-25993](https://issues.apache.org/jira/browse/SPARK-25993)
also add corresponding test cases for Parquet table.
### Why are the changes needed?
The current behavior is complex, this test case suites are designed to prevent the accidental behavior change. This pr is rebased on master, the original pr is [23108](https://github.com/apache/spark/pull/23108)
### Does this PR introduce any user-facing change?
No. This adds test cases only.
### How was this patch tested?
This is a new test case.
Closes#27130 from kevinyu98/spark-25993-2.
Authored-by: Kevin Yu <qyu@us.ibm.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
This is a followup of https://github.com/apache/spark/pull/26956 to add a migration document for 2.4.5.
### Why are the changes needed?
New legacy configuration will restore the previous behavior safely.
### Does this PR introduce any user-facing change?
This PR updates the doc.
<img width="763" alt="screenshot" src="https://user-images.githubusercontent.com/9700541/72639939-9da5a400-391b-11ea-87b1-14bca15db5a6.png">
### How was this patch tested?
Build the document and see the change manually.
Closes#27269 from dongjoon-hyun/SPARK-30312.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
[SPARK-20568](https://issues.apache.org/jira/browse/SPARK-20568) added the possibility to clean up completed files in streaming query. Deleting/archiving uses the main thread which can slow down processing. In this PR I've created thread pool to handle file delete/archival. The number of threads can be configured with `spark.sql.streaming.fileSource.cleaner.numThreads`.
### Why are the changes needed?
Do file delete/archival in separate thread.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Existing unit tests.
Closes#26502 from gaborgsomogyi/SPARK-29876.
Authored-by: Gabor Somogyi <gabor.g.somogyi@gmail.com>
Signed-off-by: Marcelo Vanzin <vanzin@cloudera.com>
The default value for backLog set back to -1, as any other value may break existing configuration by overriding Netty's default io.netty.util.NetUtil#SOMAXCONN. The documentation accordingly adjusted.
See discussion thread: https://github.com/apache/spark/pull/24732
### What changes were proposed in this pull request?
Partial rollback of https://github.com/apache/spark/pull/24732 (default for backLog set back to -1).
### Why are the changes needed?
Previous change introduces backward incompatibility by overriding default of Netty's `io.netty.util.NetUtil#SOMAXCONN`
Closes#27230 from xCASx/master.
Authored-by: Maxim Kolesnikov <swe.kolesnikov@gmail.com>
Signed-off-by: Marcelo Vanzin <vanzin@cloudera.com>
### What changes were proposed in this pull request?
SPARK-29894 provides information on the Codegen Stage Id in WEBUI for SQL Plan graphs. Similarly, this proposes to add Codegen Stage Id in the DAG visualization for Stage execution. DAGs for Stage execution are available in the WEBUI under the Jobs and Stages tabs.
### Why are the changes needed?
This is proposed as an aid for drill-down analysis of complex SQL statement execution, as it is not always easy to match parts of the SQL Plan graph with the corresponding Stage DAG execution graph. Adding Codegen Stage Id for WholeStageCodegen operations makes this task easier.
### Does this PR introduce any user-facing change?
Stage DAG visualization in the WEBUI will show codegen stage id for WholeStageCodegen operations, as in the example snippet from the WEBUI, Jobs tab (the query used in the example is TPCDS 2.4 q14a):
### How was this patch tested?
Manually tested, see also example snippet.
Closes#26675 from LucaCanali/addCodegenStageIdtoStageGraph.
Authored-by: Luca Canali <luca.canali@cern.ch>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
1) Added missing match case to SparkUncaughtExceptionHandler, so that it would not halt the process when the exception doesn't match any of the match case statements.
2) Added log message before halting process. During debugging it wasn't obvious why the Worker process would DEAD (until we set SPARK_NO_DAEMONIZE=1) due to the shell-scripts puts the process into background and essentially absorbs the exit code.
3) Added SparkUncaughtExceptionHandlerSuite. Basically we create a Spark exception-throwing application with SparkUncaughtExceptionHandler and then check its exit code.
### Why are the changes needed?
SPARK-30310, because the process would halt unexpectedly.
### How was this patch tested?
All unit tests (mvn test) were ran and OK.
Closes#26955 from tinhto-000/uncaught_exception_fix.
Authored-by: git <tinto@us.ibm.com>
Signed-off-by: Sean Owen <srowen@gmail.com>
### What changes were proposed in this pull request?
This is the second PR for the Stage Level Scheduling. This is adding in the necessary executor side changes:
1) executors to know what ResourceProfile they should be using
2) handle parsing the resource profile settings - these are not in the global configs
3) then reporting back to the driver what resource profile it was started with.
This PR adds all the piping for YARN to pass the information all the way to executors, but it just uses the default ResourceProfile (which is the global applicatino level configs).
At a high level these changes include:
1) adding a new --resourceProfileId option to the CoarseGrainedExecutorBackend
2) Add the ResourceProfile settings to new internal confs that gets passed into the Executor
3) Executor changes that use the resource profile id passed in to read the corresponding ResourceProfile confs and then parse those requests and discover resources as necessary
4) Executor registers to Driver with the Resource profile id so that the ExecutorMonitor can track how many executor with each profile are running
5) YARN side changes to show that passing the resource profile id and confs actually works. Just uses the DefaultResourceProfile for now.
I also removed a check from the CoarseGrainedExecutorBackend that used to check to make sure there were task requirements before parsing any custom resource executor requests. With the resource profiles this becomes much more expensive because we would then have to pass the task requests to each executor and the check was just a short cut and not really needed. It was much cleaner just to remove it.
Note there were some changes to the ResourceProfile, ExecutorResourceRequests, and TaskResourceRequests in this PR as well because I discovered some issues with things not being immutable. That api now look like:
val rpBuilder = new ResourceProfileBuilder()
val ereq = new ExecutorResourceRequests()
val treq = new TaskResourceRequests()
ereq.cores(2).memory("6g").memoryOverhead("2g").pysparkMemory("2g").resource("gpu", 2, "/home/tgraves/getGpus")
treq.cpus(2).resource("gpu", 2)
val resourceProfile = rpBuilder.require(ereq).require(treq).build
This makes is so that ResourceProfile is immutable and Spark can use it directly without worrying about the user changing it.
### Why are the changes needed?
These changes are needed for the executor to report which ResourceProfile they are using so that ultimately the dynamic allocation manager can use that information to know how many with a profile are running and how many more it needs to request. Its also needed to get the resource profile confs to the executor so that it can run the appropriate discovery script if needed.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Unit tests and manually on YARN.
Closes#26682 from tgravescs/SPARK-29306.
Authored-by: Thomas Graves <tgraves@nvidia.com>
Signed-off-by: Thomas Graves <tgraves@apache.org>
### What changes were proposed in this pull request?
Use the new framework to resolve the SHOW TBLPROPERTIES command. This PR along with #27243 should update all the existing V2 commands with `UnresolvedV2Relation`.
### Why are the changes needed?
This is a part of effort to make the relation lookup behavior consistent: [SPARK-2990](https://issues.apache.org/jira/browse/SPARK-29900).
### Does this PR introduce any user-facing change?
Yes `SHOW TBLPROPERTIES temp_view` now fails with `AnalysisException` will be thrown with a message `temp_view is a temp view not table`. Previously, it was returning empty row.
### How was this patch tested?
Existing tests
Closes#26921 from imback82/consistnet_v2command.
Authored-by: Terry Kim <yuminkim@gmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
Another followup of 4398dfa709
I missed two more tests added:
```
======================================================================
ERROR [0.133s]: test_to_pandas_from_mixed_dataframe (pyspark.sql.tests.test_dataframe.DataFrameTests)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/home/jenkins/python/pyspark/sql/tests/test_dataframe.py", line 617, in test_to_pandas_from_mixed_dataframe
self.assertTrue(np.all(pdf_with_only_nulls.dtypes == pdf_with_some_nulls.dtypes))
AssertionError: False is not true
======================================================================
ERROR [0.061s]: test_to_pandas_from_null_dataframe (pyspark.sql.tests.test_dataframe.DataFrameTests)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/home/jenkins/python/pyspark/sql/tests/test_dataframe.py", line 588, in test_to_pandas_from_null_dataframe
self.assertEqual(types[0], np.float64)
AssertionError: dtype('O') != <class 'numpy.float64'>
----------------------------------------------------------------------
```
### Why are the changes needed?
To make the test independent of default values of configuration.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Manually tested and Jenkins should test.
Closes#27250 from HyukjinKwon/SPARK-29188-followup2.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Add a `V1Scan` interface, so that data source v1 implementations can migrate to DS v2 much easier.
### Why are the changes needed?
It's a lot of work to migrate v1 sources to DS v2. The new API added here can allow v1 sources to go through v2 code paths without implementing all the Batch, Stream, PartitionReaderFactory, ... stuff.
We already have a v1 write fallback API after https://github.com/apache/spark/pull/25348
### Does this PR introduce any user-facing change?
no
### How was this patch tested?
new test suite
Closes#26231 from cloud-fan/v1-read-fallback.
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 explicitly disable Arrow execution for the test of toPandas empty types. If `spark.sql.execution.arrow.pyspark.enabled` is enabled by default, this test alone fails as below:
```
======================================================================
ERROR [0.205s]: test_to_pandas_from_empty_dataframe (pyspark.sql.tests.test_dataframe.DataFrameTests)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/.../pyspark/sql/tests/test_dataframe.py", line 568, in test_to_pandas_from_empty_dataframe
self.assertTrue(np.all(dtypes_when_empty_df == dtypes_when_nonempty_df))
AssertionError: False is not true
----------------------------------------------------------------------
```
it should be best to explicitly disable for the test that only works when it's disabled.
### Why are the changes needed?
To make the test independent of default values of configuration.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Manually tested and Jenkins should test.
Closes#27247 from HyukjinKwon/SPARK-29188-followup.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
In the PR, I propose to remove the SQL config `spark.sql.execution.pandas.respectSessionTimeZone` which has been deprecated since Spark 2.3.
### Why are the changes needed?
To improve code maintainability.
### Does this PR introduce any user-facing change?
Yes.
### How was this patch tested?
by running python tests, https://spark.apache.org/docs/latest/building-spark.html#pyspark-tests-with-maven-or-sbtCloses#27218 from MaxGekk/remove-respectSessionTimeZone.
Authored-by: Maxim Gekk <max.gekk@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Change ```DecisionTreeClassifier``` to ```FMClassifier``` in ```OneVsRest``` setWeightCol test
### Why are the changes needed?
In ```OneVsRest```, if the classifier doesn't support instance weight, ```OneVsRest``` weightCol will be ignored, so unit test has tested one classifier(```LogisticRegression```) that support instance weight, and one classifier (```DecisionTreeClassifier```) that doesn't support instance weight. Since ```DecisionTreeClassifier``` now supports instance weight, we need to change it to the classifier that doesn't have weight support.
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Existing test
Closes#27204 from huaxingao/spark-ovr-minor.
Authored-by: Huaxin Gao <huaxing@us.ibm.com>
Signed-off-by: zhengruifeng <ruifengz@foxmail.com>
### What changes were proposed in this pull request?
This PR aims to use `mvn` instead of `sbt` in `dev/scalastyle` to recover GitHub Action.
### Why are the changes needed?
As of now, Apache Spark sbt build is broken by the Maven Central repository policy.
https://stackoverflow.com/questions/59764749/requests-to-http-repo1-maven-org-maven2-return-a-501-https-required-status-an
> Effective January 15, 2020, The Central Maven Repository no longer supports insecure
> communication over plain HTTP and requires that all requests to the repository are
> encrypted over HTTPS.
We can reproduce this locally by the following.
```
$ rm -rf ~/.m2/repository/org/apache/apache/18/
$ build/sbt clean
```
And, in GitHub Action, `lint-scala` is the only one which is using `sbt`.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
First of all, GitHub Action should be recovered.
Also, manually, do the following.
**Without Scalastyle violation**
```
$ dev/scalastyle
OpenJDK 64-Bit Server VM warning: ignoring option MaxPermSize=384m; support was removed in 8.0
Using `mvn` from path: /usr/local/bin/mvn
Scalastyle checks passed.
```
**With Scalastyle violation**
```
$ dev/scalastyle
OpenJDK 64-Bit Server VM warning: ignoring option MaxPermSize=384m; support was removed in 8.0
Using `mvn` from path: /usr/local/bin/mvn
Scalastyle checks failed at following occurrences:
error file=/Users/dongjoon/PRS/SPARK-HTTP-501/core/src/main/scala/org/apache/spark/SparkConf.scala message=There should be no empty line separating imports in the same group. line=22 column=0
error file=/Users/dongjoon/PRS/SPARK-HTTP-501/core/src/test/scala/org/apache/spark/resource/ResourceProfileSuite.scala message=There should be no empty line separating imports in the same group. line=22 column=0
```
Closes#27242 from dongjoon-hyun/SPARK-30534.
Authored-by: Dongjoon Hyun <dhyun@apple.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
The issue here is that when Spark is downscaling the application and deletes
a few pod requests that aren't needed anymore, it may actually race with the
K8S scheduler, who may be bringing up those executors. So they may have enough
time to connect back to the driver, register, to just be deleted soon after.
This wastes resources and causes misleading entries in the driver log.
The change (ab)uses the blacklisting mechanism to consider the deleted excess
pods as blacklisted, so that if they try to connect back, the driver will deny
it.
It also changes the executor registration slightly, since even with the above
change there were misleading logs. That was because the executor registration
message was an RPC that always succeeded (bar network issues), so the executor
would always try to send an unregistration message to the driver, which would
then log several messages about not knowing anything about the executor. The
change makes the registration RPC succeed or fail directly, instead of using
the separate failure message that would lead to this issue.
Note the last change required some changes in a standalone test suite related
to dynamic allocation, since it relied on the driver not throwing exceptions
when a duplicate executor registration happened.
Tested with existing unit tests, and with live cluster with dyn alloc on.
Closes#26586 from vanzin/SPARK-29950.
Authored-by: Marcelo Vanzin <vanzin@cloudera.com>
Signed-off-by: Marcelo Vanzin <vanzin@cloudera.com>
### What changes were proposed in this pull request?
`OptimizeSkewedJoin `rule change the `outputPartitioning `after inserting `PartialShuffleReaderExec `or `SkewedPartitionReaderExec`. So it may need to introduce additional to ensure the right result. This PR disable `OptimizeSkewedJoin ` rule when introducing additional shuffle.
### Why are the changes needed?
bug fix
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
Add new ut
Closes#27226 from JkSelf/followup-skewedoptimization.
Authored-by: jiake <ke.a.jia@intel.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
TableCatalog reserves some properties, e,g `provider`, `location` for internal usage. Some of them are static once create, some of them need specific syntax to modify. Instead of using `OPTIONS (k='v')` or TBLPROPERTIES (k='v'), if k is a reserved TableCatalog property, we should use its specific syntax to add/modify/delete it. e.g. `provider` is a reserved property, we should use the `USING` clause to specify it, and should not allow `ALTER TABLE ... UNSET TBLPROPERTIES('provider')` to delete it. Also, there are two paths for v1/v2 catalog tables to resolve these properties, e.g. the v1 session catalog tables will only use the `USING` clause to decide `provider` but v2 tables will also lookup OPTION/TBLPROPERTIES(although there is a bug prohibit it).
Additionally, 'path' is not reserved but holds special meaning for `LOCATION` and it is used in `CREATE/REPLACE TABLE`'s `OPTIONS` sub-clause. Now for the session catalog tables, the `path` is case-insensitive, but for the non-session catalog tables, it is case-sensitive, we should make it both case insensitive for disambiguation.
### Why are the changes needed?
prevent reserved properties from being modified unexpectedly
unify the property resolution for v1/v2.
fix some bugs.
### Does this PR introduce any user-facing change?
yes
1 . `location` and `provider` (case sensitive) cannot be used in `CREATE/REPLACE TABLE ... OPTIONS/TBLPROPETIES` and `ALTER TABLE ... SET TBLPROPERTIES (...)`, if legacy on, they will be ignored to let the command success without having side effects
3. Once `path` in `CREATE/REPLACE TABLE ... OPTIONS` is case insensitive for v1 but sensitive for v2, but now we change it case insensitive for both kinds of tables, then v2 tables will also fail if `LOCATION` and `OPTIONS('PaTh' ='abc')` are both specified or will pick `PaTh`'s value as table location if `LOCATION` is missing.
4. Now we will detect if there are two different case `path` keys or more in `CREATE/REPLACE TABLE ... OPTIONS`, once it is a kind of unexpected last-win policy for v1, and v2 is case sensitive.
### How was this patch tested?
add ut
Closes#27197 from yaooqinn/SPARK-30507.
Authored-by: Kent Yao <yaooqinn@hotmail.com>
Signed-off-by: Wenchen Fan <wenchen@databricks.com>
### What changes were proposed in this pull request?
add setInputCol/setOutputCol in OHEModel
### Why are the changes needed?
setInputCol/setOutputCol should be in OHEModel too.
### Does this PR introduce any user-facing change?
Yes.
```OHEModel.setInputCol```
```OHEModel.setOutputCol```
### How was this patch tested?
Manually tested.
Closes#27228 from huaxingao/spark-29565.
Authored-by: Huaxin Gao <huaxing@us.ibm.com>
Signed-off-by: zhengruifeng <ruifengz@foxmail.com>
### What changes were proposed in this pull request?
In the PR, I propose to use non-deprecated constructor of `ExpressionInfo` in `SparkSessionExtensionSuite`, and pass valid strings as `examples`, `note`, `since` and `deprecated` parameters.
### Why are the changes needed?
Using another constructor allows to eliminate the following deprecation warnings while compiling Spark:
```
Warning:(335, 5) constructor ExpressionInfo in class ExpressionInfo is deprecated: see corresponding Javadoc for more information.
new ExpressionInfo("noClass", "myDb", "myFunction", "usage", "extended usage"),
Warning:(732, 5) constructor ExpressionInfo in class ExpressionInfo is deprecated: see corresponding Javadoc for more information.
new ExpressionInfo("noClass", "myDb", "myFunction2", "usage", "extended usage"),
Warning:(751, 5) constructor ExpressionInfo in class ExpressionInfo is deprecated: see corresponding Javadoc for more information.
new ExpressionInfo("noClass", "myDb", "myFunction2", "usage", "extended usage"),
```
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
By compiling and running `SparkSessionExtensionSuite`.
Closes#27221 from MaxGekk/eliminate-expr-info-warnings.
Authored-by: Maxim Gekk <max.gekk@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Enable dependency audit files to tell the value of artifact id, version, and classifier of a dependency.
For example, `avro-mapred-1.8.2-hadoop2.jar` should be expanded to `avro-mapred/1.8.2/hadoop2/avro-mapred-1.8.2-hadoop2.jar` where `avro-mapred` is the artifact id, `1.8.2` is the version, and `haddop2` is the classifier.
### Why are the changes needed?
Dependency audit files are expected to be consumed by automated tests or downstream tools.
However, current dependency audit files under `dev/deps` only show jar names. And there isn't a simple rule on how to parse the jar name to get the values of different fields. For example, `hadoop2` is the classifier of `avro-mapred-1.8.2-hadoop2.jar`, in contrast, `incubating` is the version of `htrace-core-3.1.0-incubating.jar`.
Reference: There is a good example of the downstream tool that would be enabled as yhuai suggested,
> Say we have a Spark application that depends on a third-party dependency `foo`, which pulls in `jackson` as a transient dependency. Unfortunately, `foo` depends on a different version of `jackson` than Spark. So, in the pom of this Spark application, we use the dependency management section to pin the version of `jackson`. By doing this, we are lifting `jackson` to the top-level dependency of my application and I want to have a way to keep tracking what Spark uses. What we can do is to cross-check my Spark application's classpath with what Spark uses. Then, with a test written in my code base, whenever my application bumps Spark version, this test will check what we define in the application and what Spark has, and then remind us to change our application's pom if needed. In my case, I am fine to directly access git to get these audit files.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Code changes are verified by generated dependency audit files naturally. Thus, there are no tests added.
Closes#27177 from mengCareers/depsOptimize.
Lead-authored-by: Xinrong Meng <meng.careers@gmail.com>
Co-authored-by: mengCareers <meng.careers@gmail.com>
Signed-off-by: Dongjoon Hyun <dhyun@apple.com>
### What changes were proposed in this pull request?
In the PR, I propose to support pushed down filters in CSV datasource. The reason of pushing a filter up to `UnivocityParser` is to apply the filter as soon as all its attributes become available i.e. converted from CSV fields to desired values according to the schema. This allows to skip conversions of other values if the filter returns `false`. This can improve performance when pushed filters are highly selective and conversion of CSV string fields to desired values are comparably expensive ( for example, conversion to `TIMESTAMP` values).
Here are details of the implementation:
- `UnivocityParser.convert()` converts parsed CSV tokens one-by-one sequentially starting from index 0 up to `parsedSchema.length - 1`. At current index `i`, it applies filters that refer to attributes at row fields indexes `0..i`. If any filter returns `false`, it skips conversions of other input tokens.
- Pushed filters are converted to expressions. The expressions are bound to row positions according to `requiredSchema`. The expressions are compiled to predicates via generating Java code.
- To be able to apply predicates to partially initialized rows, the predicates are grouped, and combined via the `And` expression. Final predicate at index `N` can refer to row fields at the positions `0..N`, and can be applied to a row even if other fields at the positions `N+1..requiredSchema.lenght-1` are not set.
### Why are the changes needed?
The changes improve performance on synthetic benchmarks more **than 9 times** (on JDK 8 & 11):
```
OpenJDK 64-Bit Server VM 11.0.5+10 on Mac OS X 10.15.2
Intel(R) Core(TM) i7-4850HQ CPU 2.30GHz
Filters pushdown: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
w/o filters 11889 11945 52 0.0 118893.1 1.0X
pushdown disabled 11790 11860 115 0.0 117902.3 1.0X
w/ filters 1240 1278 33 0.1 12400.8 9.6X
```
### Does this PR introduce any user-facing change?
No
### How was this patch tested?
- Added new test suite `CSVFiltersSuite`
- Added tests to `CSVSuite` and `UnivocityParserSuite`
Closes#26973 from MaxGekk/csv-filters-pushdown.
Authored-by: Maxim Gekk <max.gekk@gmail.com>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
### What changes were proposed in this pull request?
Based on the [comment](https://github.com/apache/spark/pull/26956#discussion_r366680558), this patch changes the SQL config name from `spark.sql.truncateTable.ignorePermissionAcl` to `spark.sql.truncateTable.ignorePermissionAcl.enabled`.
### Why are the changes needed?
Make this config consistent other SQL configs.
### Does this PR introduce any user-facing change?
No.
### How was this patch tested?
Unit test.
Closes#27210 from viirya/truncate-table-permission-followup.
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 followup of https://github.com/apache/spark/pull/27109. It should match the parameter lists in `createDataFrame`.
### Why are the changes needed?
To pass parameters supposed to pass.
### Does this PR introduce any user-facing change?
No (it's only in master)
### How was this patch tested?
Manually tested and existing tests should cover.
Closes#27225 from HyukjinKwon/SPARK-30434-followup.
Authored-by: HyukjinKwon <gurwls223@apache.org>
Signed-off-by: HyukjinKwon <gurwls223@apache.org>
