[SPARK-29516][SQL][TEST] Test ThriftServerQueryTestSuite asynchronously by wangyum · Pull Request #26172 · apache/spark

wangyum · 2019-10-19T05:09:12Z

What changes were proposed in this pull request?

This PR test ThriftServerQueryTestSuite in an asynchronous way.

Why are the changes needed?

The default value of spark.sql.hive.thriftServer.async is true.

Does this PR introduce any user-facing change?

No

How was this patch tested?

build/sbt "hive-thriftserver/test-only *.ThriftServerQueryTestSuite" -Phive-thriftserver
build/mvn -Dtest=none -DwildcardSuites=org.apache.spark.sql.hive.thriftserver.ThriftServerQueryTestSuite test -Phive-thriftserver

SparkQA · 2019-10-19T05:37:46Z

Test build #112309 has finished for PR 26172 at commit b3a26e1.

This patch fails Spark unit tests.
This patch merges cleanly.
This patch adds no public classes.

dongjoon-hyun · 2019-10-19T23:04:57Z

cc @zsxwing since he also asked this before.

The way I see this is that a new extended test coverage.

wangyum · 2019-10-20T01:33:44Z

It seems the reason is SQLConf.get usage in ParseDriver.scala#L91-L103. We can reproduce this issue after this commit:

build/sbt "hive-thriftserver/test-only *.ThriftServerQueryTestSuite -- -z postgreSQL" -Phive-thriftserver

The output is:

================================================================================================
statement: set spark.sql.ansi.enabled=false
spark.sql.ansi.enabled before: true
SQLConf.get.ansiEnabled: true
spark.sql.ansi.enabled after: false
================================================================================================
statement: select i, left('ahoj', i), right('ahoj', i) from range(-5, 6) t(i) order by i
spark.sql.ansi.enabled before: false
SQLConf.get.ansiEnabled: true
================================================================================================
statement: set spark.sql.ansi.enabled=true
spark.sql.ansi.enabled before: false
SQLConf.get.ansiEnabled: true
spark.sql.ansi.enabled after: true
[info] - postgreSQL/text.sql *** FAILED *** (1 second, 489 milliseconds)
[info] "-1 [info] -2 [info] -3 [info] -4 [info] -5 [info] 0 [info] 1 a j
[info] 2 ah oj
[info] 3 aho hoj
[info] 4 ahoj ahoj
[info] 5 ahoj ahoj" did not contain "Exception" Exception did not match for query #19
[info] select i, left('ahoj', i), right('ahoj', i) from range(-5, 6) t(i) order by i, expected: -1 [info] -2 [info] -3 [info] -4 [info] -5 [info] 0 [info] 1 a j
[info] 2 ah oj
[info] 3 aho hoj
[info] 4 ahoj ahoj
[info] 5 ahoj ahoj, but got: java.sql.SQLException
[info] Error running query: org.apache.spark.sql.catalyst.parser.ParseException: [info] no viable alternative at input 'left'(line 1, pos 10)
[info] [info] == SQL ==
[info] select i, left('ahoj', i), right('ahoj', i) from range(-5, 6) t(i) order by i
[info] ----------^^^ (ThriftServerQueryTestSuite.scala:225)

wangyum · 2019-10-20T01:39:33Z

cc @cloud-fan @gatorsmile @juliuszsompolski

SparkQA · 2019-10-20T01:55:04Z

Test build #112325 has finished for PR 26172 at commit 5c896ce.

This patch fails Spark unit tests.
This patch merges cleanly.
This patch adds no public classes.

AngersZhuuuu · 2019-10-21T06:15:25Z

This error happen because in SparkSession.sql() method, parse process is not under current SparkSession's conf. So some configuration about parser such as spark.sql.ansi.enable get failed in thread level

spark/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/parser/ParseDriver.scala

Line 94 in e99a9f7

lexer.legacy_setops_precedence_enbled =SQLConf.get.setOpsPrecedenceEnforced

spark/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/parser/ParseDriver.scala

Line 95 in e99a9f7

lexer.ansi =SQLConf.get.ansiEnabled

SparkQA · 2019-10-21T07:05:02Z

Test build #112367 has finished for PR 26172 at commit 8fd1efd.

This patch fails due to an unknown error code, -9.
This patch merges cleanly.
This patch adds no public classes.

wangyum · 2019-10-21T07:08:34Z

retest this please

juliuszsompolski · 2019-10-21T10:17:16Z

 * @since 2.0.0
 */
 def sql(sqlText: String): DataFrame = {
+ SparkSession.setActiveSession(this)


To keep it in scope of thriftserver, could we set it in SparkExecuteStatementOperation.execute, using the sqlContext available there?
I'm not sure what other implications it can have here, it would be safer to set there.

To keep it in scope of thriftserver, could we set it in SparkExecuteStatementOperation.execute, using the sqlContext available there?
I'm not sure what other implications it can have here, it would be safer to set there.
It's ok to set in SparkExecuteStatementOperation.execute.
#26187 (comment)
I think it ok to set this before parse sql, It was the right thing to do

Ok. Let the reviewers of #26187 decide it there. (@cloud-fan @zsxwing )

So the issue with async queries was that the parser was using the wrong dialect fallback, because it was picking the wrong SQLConf, because it was not set in the background execution thread?
@wangyum @AngersZhuuuu Does fixing this potentially fix any other of the failures in ThriftServerQueryTestSuite that are currently blacklisted to be ignored?

@juliuszsompolski Other failures can't be fixed by #26187:
spark/sql/hive-thriftserver/src/test/scala/org/apache/spark/sql/hive/thriftserver/ThriftServerQueryTestSuite.scala
Lines 87 to 102 in 7f8841b
// Missing UDF
"postgreSQL/boolean.sql",
"postgreSQL/case.sql",
// SPARK-28624
"date.sql",
// SPARK-28620
"postgreSQL/float4.sql",
// SPARK-28636
"decimalArithmeticOperations.sql",
"literals.sql",
"subquery/scalar-subquery/scalar-subquery-predicate.sql",
"subquery/in-subquery/in-limit.sql",
"subquery/in-subquery/in-group-by.sql",
"subquery/in-subquery/simple-in.sql",
"subquery/in-subquery/in-order-by.sql",
"subquery/in-subquery/in-set-operations.sql"

SparkQA · 2019-10-21T10:44:26Z

Test build #112369 has finished for PR 26172 at commit 8fd1efd.

This patch fails SparkR unit tests.
This patch merges cleanly.
This patch adds no public classes.

### What changes were proposed in this pull request? As I have comment in [SPARK-29516](#26172 (comment)) SparkSession.sql() method parse process not under current sparksession's conf, so some configuration about parser is not valid in multi-thread situation. In this pr, we add a SQLConf parameter to AbstractSqlParser and initial it with SessionState's conf. Then for each SparkSession's parser process. It will use's it's own SessionState's SQLConf and to be thread safe ### Why are the changes needed? Fix bug ### Does this PR introduce any user-facing change? NO ### How was this patch tested? NO Closes#26187 from AngersZhuuuu/SPARK-29530. Authored-by: angerszhu <angers.zhu@gmail.com> Signed-off-by: Wenchen Fan <wenchen@databricks.com>

SparkQA · 2019-10-22T03:22:02Z

Test build #112415 has finished for PR 26172 at commit 8f08b3d.

This patch passes all tests.
This patch merges cleanly.
This patch adds the following public classes (experimental):
abstract class AbstractSqlParser(conf: SQLConf) extends ParserInterface with Logging
class CatalystSqlParser(conf: SQLConf) extends AbstractSqlParser(conf)
class SparkSqlParser(conf: SQLConf) extends AbstractSqlParser(conf)

SparkQA · 2019-10-22T03:28:07Z

Test build #112416 has finished for PR 26172 at commit f291ed4.

This patch passes all tests.
This patch merges cleanly.
This patch adds no public classes.

SparkQA · 2019-10-22T03:29:51Z

Test build #112430 has finished for PR 26172 at commit 175c6f4.

This patch passes all tests.
This patch merges cleanly.
This patch adds the following public classes (experimental):
case class BitAndAgg(child: Expression) extends DeclarativeAggregate with ExpectsInputTypes
case class BitOrAgg(child: Expression) extends DeclarativeAggregate with ExpectsInputTypes
case class RepairTableStatement(tableName: Seq[String]) extends ParsedStatement

wangyum · 2019-10-22T03:47:45Z

retest this please

SparkQA · 2019-10-22T07:05:02Z

Test build #112435 has finished for PR 26172 at commit 175c6f4.

This patch fails due to an unknown error code, -9.
This patch merges cleanly.
This patch adds the following public classes (experimental):
case class BitAndAgg(child: Expression) extends DeclarativeAggregate with ExpectsInputTypes
case class BitOrAgg(child: Expression) extends DeclarativeAggregate with ExpectsInputTypes
case class RepairTableStatement(tableName: Seq[String]) extends ParsedStatement

juliuszsompolski

LGTM!

wangyum · 2019-10-22T10:21:57Z

Thank you all!

wangyum · 2019-10-22T10:22:05Z

Merged to master.

Test ThriftServerQueryTestSuite asynchronously
b3a26e1

wangyum mentioned this pull request Oct 19, 2019
[SPARK-28527][SQL][TEST] Enable ThriftServerQueryTestSuite #25868
Closed

dongjoon-hyun added the SQL label Oct 19, 2019

Add debug info
5c896ce

AngersZhuuuu mentioned this pull request Oct 21, 2019
[SPARK-29530][SQL] Make SQLConf in SQL parse process thread safe #26187
Closed

wangyum added 2 commits October 21, 2019 14:38

ApplySPARK-29530
df58bb3

Remove custom SparkConf
8fd1efd

juliuszsompolski reviewed Oct 21, 2019
View reviewed changes

wangyum added 2 commits October 22, 2019 07:21

ApplySPARK-29530
8f08b3d

Remove SparkSession.setActiveSession(this)
f291ed4

Merge remote-tracking branch 'upstream/master' intoSPARK-29516
175c6f4

wangyum changed the title ~~[WIP][SPARK-29516][SQL][TEST] Test ThriftServerQueryTestSuite asynchronously~~[SPARK-29516][SQL][TEST][test-maven] Test ThriftServerQueryTestSuite asynchronouslyOct 22, 2019

juliuszsompolski approved these changes Oct 22, 2019
View reviewed changes

wangyum changed the title ~~[SPARK-29516][SQL][TEST][test-maven] Test ThriftServerQueryTestSuite asynchronously~~[SPARK-29516][SQL][TEST] Test ThriftServerQueryTestSuite asynchronouslyOct 22, 2019

wangyum closed this in 3163b6bOct 22, 2019

wangyum deleted the SPARK-29516 branch October 22, 2019 10:21

	// Missing UDF
	"postgreSQL/boolean.sql",
	"postgreSQL/case.sql",
	// SPARK-28624
	"date.sql",
	// SPARK-28620
	"postgreSQL/float4.sql",
	// SPARK-28636
	"decimalArithmeticOperations.sql",
	"literals.sql",
	"subquery/scalar-subquery/scalar-subquery-predicate.sql",
	"subquery/in-subquery/in-limit.sql",
	"subquery/in-subquery/in-group-by.sql",
	"subquery/in-subquery/simple-in.sql",
	"subquery/in-subquery/in-order-by.sql",
	"subquery/in-subquery/in-set-operations.sql"

Uh oh!

Conversation

wangyum commented Oct 19, 2019• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What changes were proposed in this pull request?

Why are the changes needed?

Does this PR introduce any user-facing change?

How was this patch tested?

Uh oh!

SparkQA commented Oct 19, 2019

Uh oh!

dongjoon-hyun commented Oct 19, 2019

Uh oh!

wangyum commented Oct 20, 2019• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

wangyum commented Oct 20, 2019

Uh oh!

SparkQA commented Oct 20, 2019

Uh oh!

AngersZhuuuu commented Oct 21, 2019• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

SparkQA commented Oct 21, 2019

Uh oh!

wangyum commented Oct 21, 2019

Uh oh!

juliuszsompolskiOct 21, 2019

Choose a reason for hiding this comment

Uh oh!

AngersZhuuuuOct 21, 2019

Choose a reason for hiding this comment

Uh oh!

juliuszsompolskiOct 21, 2019

Choose a reason for hiding this comment

Uh oh!

juliuszsompolskiOct 21, 2019

Choose a reason for hiding this comment

Uh oh!

wangyumOct 22, 2019

Choose a reason for hiding this comment

Uh oh!

SparkQA commented Oct 21, 2019

Uh oh!

SparkQA commented Oct 22, 2019

Uh oh!

SparkQA commented Oct 22, 2019

Uh oh!

SparkQA commented Oct 22, 2019

Uh oh!

wangyum commented Oct 22, 2019

Uh oh!

SparkQA commented Oct 22, 2019

Uh oh!

juliuszsompolski left a comment

Choose a reason for hiding this comment

Uh oh!

wangyum commented Oct 22, 2019

Uh oh!

wangyum commented Oct 22, 2019

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

wangyum commented Oct 19, 2019•
edited
Loading

wangyum commented Oct 20, 2019•
edited
Loading

AngersZhuuuu commented Oct 21, 2019•
edited
Loading