Skip to content

[SPARK-39961][SQL] DS V2 push-down translate Cast if the cast is safe #37388

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
beliefer wants to merge 2 commits into from
Closed

[SPARK-39961][SQL] DS V2 push-down translate Cast if the cast is safe #37388

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b3c0720
[SPARK-39961][SQL] DS V2 push-down translate Cast if the cast is safe
beliefer Aug 3, 2022
feb2a13
Update code
beliefer Aug 4, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion sql/core/src/main/scala/org/apache/spark/sql/catalyst/util/V2ExpressionBuilder.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,8 @@ class V2ExpressionBuilder(e: Expression, isPredicate: Boolean = false) {
} else {
None
}
case Cast(child, dataType, _, true) =>
case Cast(child, dataType, _, ansiEnabled)
if Cast.canUpCast(child.dataType, dataType) || ansiEnabled =>
Copy link
Contributor

@huaxingao huaxingao Aug 3, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit: 4-space indentation?

Copy link
Member

@dongjoon-hyun dongjoon-hyun Aug 3, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

+1 for @huaxingao 's comment.

In addition, let's switch the predicate order, @beliefer .

- if Cast.canUpCast(child.dataType, dataType) || ansiEnabled =>
+ if ansiEnabled || Cast.canUpCast(child.dataType, dataType) =>

Copy link
Contributor Author

@beliefer beliefer Aug 4, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Got it.

generateExpression(child).map(v => new V2Cast(v, dataType))
case Abs(child, true) => generateExpressionWithName("ABS", Seq(child))
case Coalesce(children) => generateExpressionWithName("COALESCE", children)
Expand Down
45 changes: 18 additions & 27 deletions sql/core/src/test/scala/org/apache/spark/sql/jdbc/JDBCV2Suite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1109,7 +1109,7 @@ class JDBCV2Suite extends QueryTest with SharedSparkSession with ExplainSuiteHel
"CAST(BONUS AS string) LIKE '%30%', CAST(DEPT AS byte) > 1, " +
"CAST(DEPT AS short) > 1, CAST(BONUS AS decimal(20,2)) > 1200.00]"
} else {
"PushedFilters: [BONUS IS NOT NULL, DEPT IS NOT NULL],"
"PushedFilters: [BONUS IS NOT NULL, DEPT IS NOT NULL, CAST(BONUS AS string) LIKE '%30%']"
}
checkPushedInfo(df6, expectedPlanFragment6)
checkAnswer(df6, Seq(Row(2, "david", 10000, 1300, true)))
Expand Down Expand Up @@ -1199,18 +1199,16 @@ class JDBCV2Suite extends QueryTest with SharedSparkSession with ExplainSuiteHel
checkPushedInfo(df1, "PushedFilters: [CHAR_LENGTH(NAME) > 2],")
checkAnswer(df1, Seq(Row("fred", 1), Row("mary", 2)))

withSQLConf(SQLConf.ANSI_ENABLED.key -> "true") {
val df2 = sql(
"""
|SELECT *
|FROM h2.test.people
|WHERE h2.my_strlen(CASE WHEN NAME = 'fred' THEN NAME ELSE "abc" END) > 2
val df2 = sql(
"""
|SELECT *
|FROM h2.test.people
|WHERE h2.my_strlen(CASE WHEN NAME = 'fred' THEN NAME ELSE "abc" END) > 2
""".stripMargin)
checkFiltersRemoved(df2)
checkPushedInfo(df2,
"PushedFilters: [CHAR_LENGTH(CASE WHEN NAME = 'fred' THEN NAME ELSE 'abc' END) > 2],")
checkAnswer(df2, Seq(Row("fred", 1), Row("mary", 2)))
}
checkFiltersRemoved(df2)
checkPushedInfo(df2,
"PushedFilters: [CHAR_LENGTH(CASE WHEN NAME = 'fred' THEN NAME ELSE 'abc' END) > 2],")
checkAnswer(df2, Seq(Row("fred", 1), Row("mary", 2)))
} finally {
JdbcDialects.unregisterDialect(testH2Dialect)
JdbcDialects.registerDialect(H2Dialect)
Expand Down Expand Up @@ -2262,24 +2260,17 @@ class JDBCV2Suite extends QueryTest with SharedSparkSession with ExplainSuiteHel
}

test("scan with aggregate push-down: partial push-down AVG with overflow") {
def createDataFrame: DataFrame = spark.read
.option("partitionColumn", "id")
.option("lowerBound", "0")
.option("upperBound", "2")
.option("numPartitions", "2")
.table("h2.test.item")
.agg(avg($"PRICE").as("avg"))

Seq(true, false).foreach { ansiEnabled =>
withSQLConf((SQLConf.ANSI_ENABLED.key, ansiEnabled.toString)) {
val df = createDataFrame
val df = spark.read
.option("partitionColumn", "id")
.option("lowerBound", "0")
.option("upperBound", "2")
.option("numPartitions", "2")
.table("h2.test.item")
.agg(avg($"PRICE").as("avg"))
checkAggregateRemoved(df, false)
df.queryExecution.optimizedPlan.collect {
case _: DataSourceV2ScanRelation =>
val expected_plan_fragment =
"PushedAggregates: [COUNT(PRICE), SUM(PRICE)]"
checkKeywordsExistsInExplain(df, expected_plan_fragment)
}
checkPushedInfo(df, "PushedAggregates: [COUNT(PRICE), SUM(PRICE)]")
if (ansiEnabled) {
val e = intercept[SparkException] {
df.collect()
Expand Down