ODIn/spark-instrumented-optimizer
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

[SPARK-16639][SQL] The query with having condition that contains grouping by column should work

Browse source 
## What changes were proposed in this pull request?

The query with having condition that contains grouping by column will be failed during analysis. E.g.,

    create table tbl(a int, b string);
    select count(b) from tbl group by a + 1 having a + 1 = 2;

Having condition should be able to use grouping by column.

## How was this patch tested?

Jenkins tests.

Author: Liang-Chi Hsieh <simonh@tw.ibm.com>

Closes #14296 from viirya/having-contains-grouping-column.
This commit is contained in:
Liang-Chi Hsieh 2016-07-28 22:33:33 +08:00 committed by Wenchen Fan
parent 762366fd87
commit 9ade77c3fa
2 changed files with 31 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

15
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala
View file

@ -246,7 +246,7 @@ class Analyzer(
}.isDefined
}
private def hasGroupingFunction(e: Expression): Boolean = {
private[sql] def hasGroupingFunction(e: Expression): Boolean = {
e.collectFirst {
case g: Grouping => g
case g: GroupingID => g
@ -1207,6 +1207,19 @@ class Analyzer(
val alias = Alias(ae, ae.toString)()
aggregateExpressions += alias
alias.toAttribute
// Grouping functions are handled in the rule [[ResolveGroupingAnalytics]].
case e: Expression if grouping.exists(_.semanticEquals(e)) &&
!ResolveGroupingAnalytics.hasGroupingFunction(e) &&
!aggregate.output.exists(_.semanticEquals(e)) =>
e match {
case ne: NamedExpression =>
aggregateExpressions += ne
ne.toAttribute
case _ =>
val alias = Alias(e, e.toString)()
aggregateExpressions += alias
alias.toAttribute
}
}
// Push the aggregate expressions into the aggregate (if any).

12
sql/core/src/test/scala/org/apache/spark/sql/SQLQuerySuite.scala
View file

@ -39,12 +39,24 @@ class SQLQuerySuite extends QueryTest with SharedSQLContext {
setupTestData()
test("having clause") {
withTempView("hav") {
Seq(("one", 1), ("two", 2), ("three", 3), ("one", 5)).toDF("k", "v")
.createOrReplaceTempView("hav")
checkAnswer(
sql("SELECT k, sum(v) FROM hav GROUP BY k HAVING sum(v) > 2"),
Row("one", 6) :: Row("three", 3) :: Nil)
}
}
test("having condition contains grouping column") {
withTempView("hav") {
Seq(("one", 1), ("two", 2), ("three", 3), ("one", 5)).toDF("k", "v")
.createOrReplaceTempView("hav")
checkAnswer(
sql("SELECT count(k) FROM hav GROUP BY v + 1 HAVING v + 1 = 2"),
Row(1) :: Nil)
}
}
test("SPARK-8010: promote numeric to string") {
val df = Seq((1, 1)).toDF("key", "value")