[SPARK-25121][SQL] Supports multi-part relation names for join strategy hint resolution #27935

maropu · 2020-03-17T07:38:54Z

What changes were proposed in this pull request?

This pr fixed code for respecting a multi-part identifier (e.g.,dbname.tablename) for join strategy hint resolution. For example, the master ignores a database name in a hint parameter;

scala> sql("CREATE DATABASE testDb")
scala> spark.range(10).write.saveAsTable("testDb.t")

// without this patch
scala> spark.range(10).join(spark.table("testDb.t"), "id").hint("broadcast", "testDb.t").explain
== Physical Plan ==
*(2) Project [id#24L]
+- *(2) BroadcastHashJoin [id#24L], [id#26L], Inner, BuildLeft
   :- BroadcastExchange HashedRelationBroadcastMode(List(input[0, bigint, false]))
   :  +- *(1) Range (0, 10, step=1, splits=4)
   +- *(2) Project [id#26L]
      +- *(2) Filter isnotnull(id#26L)
         +- *(2) FileScan parquet testdb.t[id#26L] Batched: true, Format: Parquet, Location: InMemoryFileIndex[file:/Users/maropu/Repositories/spark/spark-2.3.1-bin-hadoop2.7/spark-warehouse..., PartitionFilters: [], PushedFilters: [IsNotNull(id)], ReadSchema: struct<id:bigint>

// with this patch
scala> spark.range(10).join(spark.table("testDb.t"), "id").hint("broadcast", "testDb.t").explain
== Physical Plan ==
*(2) Project [id#3L]
+- *(2) BroadcastHashJoin [id#3L], [id#5L], Inner, BuildRight
   :- *(2) Range (0, 10, step=1, splits=4)
   +- BroadcastExchange HashedRelationBroadcastMode(List(input[0, bigint, true]))
      +- *(1) Project [id#5L]
         +- *(1) Filter isnotnull(id#5L)
            +- *(1) FileScan parquet testdb.t[id#5L] Batched: true, Format: Parquet, Location: InMemoryFileIndex[file:/Users/maropu/Repositories/spark/spark-master/spark-warehouse/testdb.db/t], PartitionFilters: [], PushedFilters: [IsNotNull(id)], ReadSchema: struct<id:bigint>

This PR comes from #22198

Why are the changes needed?

For better usability.

Does this PR introduce any user-facing change?

No.

How was this patch tested?

Added unit tests.

maropu · 2020-03-17T07:39:30Z

Not ready for reviews.

SparkQA · 2020-03-17T09:44:17Z

Test build #119924 has finished for PR 27935 at commit 57ced2b.

This patch fails Spark unit tests.
This patch merges cleanly.
This patch adds the following public classes (experimental):
trait BaseIdentifier
sealed trait IdentifierWithDatabase extends BaseIdentifier
case class AliasIdentifier(name: String, qualifier: Seq[String]) extends BaseIdentifier

SparkQA · 2020-03-17T15:34:36Z

Test build #119929 has finished for PR 27935 at commit b276c1c.

This patch passes all tests.
This patch merges cleanly.
This patch adds the following public classes (experimental):
trait BaseIdentifier
sealed trait IdentifierWithDatabase extends BaseIdentifier
case class AliasIdentifier(name: String, qualifier: Seq[String]) extends BaseIdentifier

SparkQA · 2020-03-17T17:30:55Z

Test build #119931 has finished for PR 27935 at commit 9a0c68c.

This patch passes all tests.
This patch merges cleanly.
This patch adds the following public classes (experimental):
trait BaseIdentifier
sealed trait IdentifierWithDatabase extends BaseIdentifier
case class AliasIdentifier(name: String, qualifier: Seq[String]) extends BaseIdentifier

maropu · 2020-03-18T04:41:00Z

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveHints.scala

+    //     For example, in a query `SELECT /* BROADCAST(default.t) */ * FROM default.t JOIN t`,
+    //     the broadcast hint will match the left-side table only, `default.t`.
+    //
+    //  3. otherwise, no match happens.


I re-read your comments again in #22198 and summarized up them above. If I misunderstand something, please let me know. @dongjoon-hyun @cloud-fan

SparkQA · 2020-03-18T07:05:02Z

Test build #119963 has finished for PR 27935 at commit 52089fc.

This patch fails due to an unknown error code, -9.
This patch merges cleanly.
This patch adds the following public classes (experimental):
trait BaseIdentifier
sealed trait IdentifierWithDatabase extends BaseIdentifier
case class AliasIdentifier(name: String, qualifier: Seq[String]) extends BaseIdentifier

cloud-fan · 2020-03-18T08:25:45Z

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveHints.scala

+    //
+    //  1. they match if an identifier in a hint only has one part and it is the same with
+    //     a relation name in a query. If a relation has a namespace (`db1.t`), we just ignore it.
+    //     For example, in a query `SELECT /* BROADCAST(t) */ * FROM db1.t JOIN t`,


is this the existing behavior?

cc @maryannxue as well

Yea, I checked queries below in v2.4.5;

// v2.4.5 scala> sql("CREATE DATABASE db1") scala> sql("CREATE TABLE db1.t(key int)") scala> sql("CREATE TABLE t(key int)") scala> sql("""SELECT /*+ MAPJOIN(t) */ * FROM db1.t JOIN t""") == Parsed Logical Plan == 'UnresolvedHint MAPJOIN, ['t] +- 'Project [*] +- 'Join Inner :- 'UnresolvedRelation `db1`.`t` +- 'UnresolvedRelation `t` == Analyzed Logical Plan == key: int, key: int Project [key#20, key#21] +- Join Inner :- ResolvedHint (broadcast) : +- SubqueryAlias `db1`.`t` : +- HiveTableRelation `db1`.`t`, org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe, [key#20] +- ResolvedHint (broadcast) +- SubqueryAlias `default`.`t` +- HiveTableRelation `default`.`t`, org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe, [key#21]

cloud-fan · 2020-03-18T14:47:22Z

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveHints.scala

+    //     For example, in a query `SELECT /*+ BROADCAST(t) */ * FROM db1.t JOIN t`,
+    //     the broadcast hint will match both tables, `db1.t` and `t`.
+    //
+    //  2. they match if an identifier in a hint has two parts and it is the same with


What if the identifier has more than 2 parts like cata.ns1.ns2.tbl ? How about we define a simple rule: If identInHint is a tail of identInQuery?

Ah, that looks nice. I'll update.

How about the latest update?

SparkQA · 2020-03-18T16:34:16Z

Test build #119987 has finished for PR 27935 at commit ac19ea1.

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

SparkQA · 2020-03-19T04:06:47Z

Test build #120012 has finished for PR 27935 at commit 5a0b4ed.

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

SparkQA · 2020-03-19T04:11:53Z

Test build #120011 has finished for PR 27935 at commit 4628aa0.

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

SparkQA · 2020-03-19T04:11:56Z

Test build #120013 has finished for PR 27935 at commit 70c994a.

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

cloud-fan · 2020-03-19T06:22:53Z

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/identifiers.scala


 package org.apache.spark.sql.catalyst

+trait BaseIdentifier {


why do we add a base trait?

Ah, we can remove it now. I'll update.

cloud-fan · 2020-03-19T06:26:56Z

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveHints.scala

        plan: LogicalPlan,
-        relations: mutable.HashSet[String],
+        relationsInHint: Seq[Seq[String]],
+        appliedRelations: mutable.ArrayBuffer[Seq[String]],


I think we don't care which relations are matched, but which relation name specified by hint does not have a match.

how about relationsInHintWithMatch: mutable.HashSet[Seq[String]]

and in code

relationsInHint.find(matchedIdentifier(_, ident)).map { relation => relationsInHintWithMatch += relation plan with hint applied }.getOrElse { originalPlan }

~~But, without this variable, how do we track non-used hints for the error report in hintErrorHandler.hintRelationsNotFound?~~

updated in the latest commit.

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveHints.scala

cloud-fan · 2020-03-19T08:47:36Z

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveHints.scala

-              if relations.exists(resolver(_, ident.last)) =>
-            relations.remove(ident.last)
+              if relationsInHint.exists(matchedIdentifier(_, ident)) =>
+            relationsInHintWithMatch += ident


The problem here is, ident is the actual relation name, not the relation name in the hint. This forces us to do an extra case insensitive match in https://github.com/apache/spark/pull/27935/files#diff-746a6d090224c7cfbe15daa27fa27408R163

Urr, I see. I'll fix that.

Since your suggested code above broke some existing tests, I modified it a little based on that. Updated in the latest commit.

cloud-fan · 2020-03-19T13:22:17Z

sql/core/src/test/scala/org/apache/spark/sql/execution/SQLViewSuite.scala

    }
  }
+
+  test("broadcast hint on temp view") {


Seems this test suite is only for permanent views. Maybe we can put everything in one test in https://github.com/apache/spark/pull/27935/files#diff-fa1d044f9cfe587e27866393fe18fd46R329

Moved this test into DataFrameJoinSuite.

cloud-fan

LGTM except one comment

SparkQA · 2020-03-19T14:02:04Z

Test build #120043 has finished for PR 27935 at commit f07613d.

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

SparkQA · 2020-03-19T17:50:47Z

Test build #120052 has finished for PR 27935 at commit 6528dad.

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

SparkQA · 2020-03-19T18:38:25Z

Test build #120056 has finished for PR 27935 at commit c517c1f.

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

dongjoon-hyun · 2020-03-19T21:21:42Z

sql/core/src/test/scala/org/apache/spark/sql/DataFrameJoinSuite.scala

+          val plan = sql(s"SELECT * FROM $dbName.$table1Name, $dbName.$table2Name " +
+            s"WHERE $table1Name.id = $table2Name.id")
+            .queryExecution.executedPlan
+          assert(plan.collect { case p: BroadcastHashJoinExec => p }.isEmpty)


checkIfHintNotApplied?

- .queryExecution.executedPlan - assert(plan.collect { case p: BroadcastHashJoinExec => p }.isEmpty) + checkIfHintNotApplied(plan)

Ah... I forgot to remove the old tests... I can remove it. Thanks!

dongjoon-hyun · 2020-03-19T21:40:37Z

sql/core/src/test/scala/org/apache/spark/sql/DataFrameJoinSuite.scala

+          checkIfHintApplied(sqlTemplate(s"$dbName.$table1Name", s"$dbName.$table1Name"))
+          checkIfHintApplied(sqlTemplate(s"$dbName.$table1Name", table1Name))
+          checkIfHintNotApplied(sqlTemplate(table1Name, s"$dbName.$table1Name"))
+          checkIfHintNotApplied(sqlTemplate(s"$dbName.$table1Name", s"$dbName.$table1Name.id"))


Is $dbName.$table1Name.id used as a negative value for hintTableName? It's a little confusing because there is a catalog concept. For three fields, can we use catalog instead of id?

ok, I will modify id -> spark_catalog.

dongjoon-hyun · 2020-03-19T21:58:12Z

sql/core/src/test/scala/org/apache/spark/sql/DataFrameJoinSuite.scala

+          withTempView("tv") {
+            sql(s"CREATE TEMPORARY VIEW tv AS SELECT * FROM $dbName.$table1Name")
+            checkIfHintApplied(sqlTemplate("tv", "tv"))
+            checkIfHintNotApplied(sqlTemplate("tv", "default.tv"))


This might be misleading. Technically, this is the same with sqlTemplate("tv", "non_exist") because we cannot use a database qualifier for Temporary View.

scala> sql("select * from default.tv").show org.apache.spark.sql.AnalysisException: Table or view not found: default.tv; line 1 pos 14;

Yea, but I think a query with a hint having a non-existent relation identifier should work?;

scala> sql("create table t1 (id int)") scala> sql("create table t2 (id int)") scala> sql("create temporary view tv as select * from t1") scala> sql("SELECT /*+ BROADCASTJOIN(default.non_exist) */ * FROM tv, t2 WHERE tv.id = t2.id").explain(true) 20/03/20 07:34:02 WARN HintErrorLogger: Count not find relation 'default.non_exist' specified in hint 'BROADCASTJOIN(default.non_exist)'. == Parsed Logical Plan == 'UnresolvedHint BROADCASTJOIN, ['default.non_exist] +- 'Project [*] +- 'Filter ('tv.id = 't2.id) +- 'Join Inner :- 'UnresolvedRelation [tv] +- 'UnresolvedRelation [t2] == Analyzed Logical Plan == id: int, id: int Project [id#0, id#1] +- Filter (id#0 = id#1) +- Join Inner :- SubqueryAlias tv : +- Project [id#0] : +- SubqueryAlias spark_catalog.default.t1 : +- Relation[id#0] parquet +- SubqueryAlias spark_catalog.default.t2 +- Relation[id#1] parquet

I will modify the tests a little.

dongjoon-hyun · 2020-03-19T22:07:19Z

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveHints.scala

+    //
+    // For example,
+    //  * in a query `SELECT /*+ BROADCAST(t) */ * FROM db1.t JOIN t`,
+    //    the broadcast hint will match both tables, `db1.t` and `t`.


- the broadcast hint will match both tables, `db1.t` and `t`. + the broadcast hint will match both tables, `db1.t` and `t`, even when the current db is `db2`.

dongjoon-hyun · 2020-03-19T22:19:53Z

sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/analysis/ResolveHintsSuite.scala

+    // local temp table (single-part identifier case)
+    checkAnalysis(
+      UnresolvedHint("MAPJOIN", Seq("table", "table2"),
+        table("table").join(table("table2"))),


The following will be better because this is a caseSensitive = false test case.

- table("table").join(table("table2"))), + table("TaBlE").join(table("TaBlE2"))),

maropu · 2020-03-19T23:03:16Z

Thanks for the reviews, @dongjoon-hyun! I've updated, so could you check the latest commit?

dongjoon-hyun · 2020-03-19T23:28:48Z

sql/core/src/test/scala/org/apache/spark/sql/DataFrameJoinSuite.scala

-          val plan = sql(s"SELECT * FROM $dbName.$table1Name, $dbName.$table2Name " +
-            s"WHERE $table1Name.id = $table2Name.id")
-            .queryExecution.executedPlan
-          assert(plan.collect { case p: BroadcastHashJoinExec => p }.isEmpty)


Oh, is this pre-testing removed completely?

Ah, I saw that your comment.

[SPARK-25121][SQL] Supports multi-part relation names for join strategy hint resolution #27935 (comment)

Yea, I think the current test coverage is good enough.

SparkQA · 2020-03-20T03:09:18Z

Test build #120074 has finished for PR 27935 at commit 61e4a95.

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

dongjoon-hyun

+1, LGTM.

dongjoon-hyun · 2020-03-20T03:10:28Z

Merged to master/3.0. Thank you, @maropu and @cloud-fan .

… resolution ### What changes were proposed in this pull request? This pr fixed code to respect a database name for broadcast table hint resolution. Currently, spark ignores a database name in multi-part names; ``` scala> sql("CREATE DATABASE testDb") scala> spark.range(10).write.saveAsTable("testDb.t") // without this patch scala> spark.range(10).join(spark.table("testDb.t"), "id").hint("broadcast", "testDb.t").explain == Physical Plan == *(2) Project [id#24L] +- *(2) BroadcastHashJoin [id#24L], [id#26L], Inner, BuildLeft :- BroadcastExchange HashedRelationBroadcastMode(List(input[0, bigint, false])) : +- *(1) Range (0, 10, step=1, splits=4) +- *(2) Project [id#26L] +- *(2) Filter isnotnull(id#26L) +- *(2) FileScan parquet testdb.t[id#26L] Batched: true, Format: Parquet, Location: InMemoryFileIndex[file:/Users/maropu/Repositories/spark/spark-2.3.1-bin-hadoop2.7/spark-warehouse..., PartitionFilters: [], PushedFilters: [IsNotNull(id)], ReadSchema: struct<id:bigint> // with this patch scala> spark.range(10).join(spark.table("testDb.t"), "id").hint("broadcast", "testDb.t").explain == Physical Plan == *(2) Project [id#3L] +- *(2) BroadcastHashJoin [id#3L], [id#5L], Inner, BuildRight :- *(2) Range (0, 10, step=1, splits=4) +- BroadcastExchange HashedRelationBroadcastMode(List(input[0, bigint, true])) +- *(1) Project [id#5L] +- *(1) Filter isnotnull(id#5L) +- *(1) FileScan parquet testdb.t[id#5L] Batched: true, Format: Parquet, Location: InMemoryFileIndex[file:/Users/maropu/Repositories/spark/spark-master/spark-warehouse/testdb.db/t], PartitionFilters: [], PushedFilters: [IsNotNull(id)], ReadSchema: struct<id:bigint> ``` This PR comes from #22198 ### Why are the changes needed? For better usability. ### Does this PR introduce any user-facing change? No. ### How was this patch tested? Added unit tests. Closes #27935 from maropu/SPARK-25121-2. Authored-by: Takeshi Yamamuro <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]> (cherry picked from commit ca499e9) Signed-off-by: Dongjoon Hyun <[email protected]>

maropu · 2020-03-20T05:26:02Z

Thanks for the reviews, @dongjoon-hyun @cloud-fan !

gatorsmile · 2020-03-25T02:16:36Z

@maropu Could you fix the title and PR description? Add the test cases for the other join hints? Thanks!

maropu · 2020-03-25T05:22:15Z

Sure, I'll do that. Thanks!

…ifiers in join strategy hints ### What changes were proposed in this pull request? This pr intends to add unit tests for the other join hints (`MERGEJOIN`, `SHUFFLE_HASH`, and `SHUFFLE_REPLICATE_NL`). This is a followup PR of #27935. ### Why are the changes needed? For better test coverage. ### Does this PR introduce any user-facing change? No. ### How was this patch tested? Added unit tests. Closes #28013 from maropu/SPARK-25121-FOLLOWUP. Authored-by: Takeshi Yamamuro <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>

…ifiers in join strategy hints ### What changes were proposed in this pull request? This pr intends to add unit tests for the other join hints (`MERGEJOIN`, `SHUFFLE_HASH`, and `SHUFFLE_REPLICATE_NL`). This is a followup PR of #27935. ### Why are the changes needed? For better test coverage. ### Does this PR introduce any user-facing change? No. ### How was this patch tested? Added unit tests. Closes #28013 from maropu/SPARK-25121-FOLLOWUP. Authored-by: Takeshi Yamamuro <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]> (cherry picked from commit da49f50) Signed-off-by: Dongjoon Hyun <[email protected]>

… resolution ### What changes were proposed in this pull request? This pr fixed code to respect a database name for broadcast table hint resolution. Currently, spark ignores a database name in multi-part names; ``` scala> sql("CREATE DATABASE testDb") scala> spark.range(10).write.saveAsTable("testDb.t") // without this patch scala> spark.range(10).join(spark.table("testDb.t"), "id").hint("broadcast", "testDb.t").explain == Physical Plan == *(2) Project [id#24L] +- *(2) BroadcastHashJoin [id#24L], [id#26L], Inner, BuildLeft :- BroadcastExchange HashedRelationBroadcastMode(List(input[0, bigint, false])) : +- *(1) Range (0, 10, step=1, splits=4) +- *(2) Project [id#26L] +- *(2) Filter isnotnull(id#26L) +- *(2) FileScan parquet testdb.t[id#26L] Batched: true, Format: Parquet, Location: InMemoryFileIndex[file:/Users/maropu/Repositories/spark/spark-2.3.1-bin-hadoop2.7/spark-warehouse..., PartitionFilters: [], PushedFilters: [IsNotNull(id)], ReadSchema: struct<id:bigint> // with this patch scala> spark.range(10).join(spark.table("testDb.t"), "id").hint("broadcast", "testDb.t").explain == Physical Plan == *(2) Project [id#3L] +- *(2) BroadcastHashJoin [id#3L], [id#5L], Inner, BuildRight :- *(2) Range (0, 10, step=1, splits=4) +- BroadcastExchange HashedRelationBroadcastMode(List(input[0, bigint, true])) +- *(1) Project [id#5L] +- *(1) Filter isnotnull(id#5L) +- *(1) FileScan parquet testdb.t[id#5L] Batched: true, Format: Parquet, Location: InMemoryFileIndex[file:/Users/maropu/Repositories/spark/spark-master/spark-warehouse/testdb.db/t], PartitionFilters: [], PushedFilters: [IsNotNull(id)], ReadSchema: struct<id:bigint> ``` This PR comes from apache#22198 ### Why are the changes needed? For better usability. ### Does this PR introduce any user-facing change? No. ### How was this patch tested? Added unit tests. Closes apache#27935 from maropu/SPARK-25121-2. Authored-by: Takeshi Yamamuro <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>

…ifiers in join strategy hints ### What changes were proposed in this pull request? This pr intends to add unit tests for the other join hints (`MERGEJOIN`, `SHUFFLE_HASH`, and `SHUFFLE_REPLICATE_NL`). This is a followup PR of apache#27935. ### Why are the changes needed? For better test coverage. ### Does this PR introduce any user-facing change? No. ### How was this patch tested? Added unit tests. Closes apache#28013 from maropu/SPARK-25121-FOLLOWUP. Authored-by: Takeshi Yamamuro <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>

maropu force-pushed the SPARK-25121-2 branch 2 times, most recently from b276c1c to 9a0c68c Compare March 17, 2020 12:18

maropu force-pushed the SPARK-25121-2 branch from 9a0c68c to 2ce7d7a Compare March 18, 2020 04:29

Fix

52089fc

maropu force-pushed the SPARK-25121-2 branch from 2ce7d7a to 52089fc Compare March 18, 2020 04:36

maropu changed the title ~~[WIP][SPARK-25121][SQL] Supports multi-part table names for broadcast hint resolution~~ [SPARK-25121][SQL] Supports multi-part table names for broadcast hint resolution Mar 18, 2020

maropu commented Mar 18, 2020

View reviewed changes

cloud-fan reviewed Mar 18, 2020

View reviewed changes

Fix

ac19ea1

cloud-fan reviewed Mar 18, 2020

View reviewed changes

maropu force-pushed the SPARK-25121-2 branch from 4628aa0 to 5a0b4ed Compare March 18, 2020 23:32

Fix

70c994a

maropu force-pushed the SPARK-25121-2 branch from 5a0b4ed to 70c994a Compare March 18, 2020 23:44

cloud-fan reviewed Mar 19, 2020

View reviewed changes

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveHints.scala Show resolved Hide resolved

Fix

fbbe824

maropu force-pushed the SPARK-25121-2 branch from f54fe77 to fbbe824 Compare March 19, 2020 07:31

Fix

2444a20

cloud-fan reviewed Mar 19, 2020

View reviewed changes

Fix

f07613d

Fix

6528dad

cloud-fan reviewed Mar 19, 2020

View reviewed changes

cloud-fan approved these changes Mar 19, 2020

View reviewed changes

Fix

c517c1f

dongjoon-hyun added the SQL label Mar 19, 2020

dongjoon-hyun reviewed Mar 19, 2020

View reviewed changes

Address Dongjoon reviews

61e4a95

dongjoon-hyun reviewed Mar 19, 2020

View reviewed changes

dongjoon-hyun approved these changes Mar 20, 2020

View reviewed changes

dongjoon-hyun closed this in ca499e9 Mar 20, 2020

maropu changed the title ~~[SPARK-25121][SQL] Supports multi-part table names for broadcast hint resolution~~ [SPARK-25121][SQL] Supports multi-part relation names for join strategy hint resolution Mar 25, 2020

maropu mentioned this pull request Mar 25, 2020

[SPARK-25121][SQL][FOLLOWUP] Add more unit tests for multi-part identifiers in join strategy hints #28013

Closed

[SPARK-25121][SQL] Supports multi-part relation names for join strategy hint resolution #27935

[SPARK-25121][SQL] Supports multi-part relation names for join strategy hint resolution #27935

Uh oh!

Conversation

maropu commented Mar 17, 2020 • 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!

maropu commented Mar 17, 2020

Uh oh!

SparkQA commented Mar 17, 2020

Uh oh!

SparkQA commented Mar 17, 2020

Uh oh!

SparkQA commented Mar 17, 2020

Uh oh!

maropu Mar 18, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

SparkQA commented Mar 18, 2020

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

maropu Mar 18, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

SparkQA commented Mar 18, 2020

Uh oh!

SparkQA commented Mar 19, 2020

Uh oh!

SparkQA commented Mar 19, 2020

Uh oh!

SparkQA commented Mar 19, 2020

Uh oh!

Choose a reason for hiding this comment

Uh oh!

maropu Mar 19, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

maropu Mar 19, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

maropu Mar 19, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cloud-fan Mar 19, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

maropu commented Mar 17, 2020 •

edited

Loading

maropu Mar 18, 2020 •

edited

Loading

maropu Mar 18, 2020 •

edited

Loading

maropu Mar 19, 2020 •

edited

Loading

maropu Mar 19, 2020 •

edited

Loading

maropu Mar 19, 2020 •

edited

Loading

cloud-fan Mar 19, 2020 •

edited

Loading

dongjoon-hyun Mar 19, 2020 •

edited

Loading