viirya commented on a change in pull request #27499: [SPARK-30590][SQL] Untyped
select API cannot take typed column expression that needs input type
URL: https://github.com/apache/spark/pull/27499#discussion_r384122068
##########
File path:
sql/core/src/test/scala/org/apache/spark/sql/DatasetAggregatorSuite.scala
##########
@@ -394,4 +403,19 @@ class DatasetAggregatorSuite extends QueryTest with
SharedSparkSession {
checkAnswer(group, Row("bob", Row(true, 3)) :: Nil)
checkDataset(group.as[OptionBooleanIntData], OptionBooleanIntData("bob",
Some((true, 3))))
}
+
+ test("SPARK-30590: untyped select should not accept typed column without
input type") {
+ val df = Seq((1, 2, 3, 4, 5, 6)).toDF("a", "b", "c", "d", "e", "f")
+ val fooAgg = (i: Int) => FooAgg(i).toColumn.name(s"foo_agg_$i")
+
+ val agg1 = df.select(fooAgg(1), fooAgg(2), fooAgg(3), fooAgg(4), fooAgg(5))
+ checkDataset(agg1, (3, 5, 7, 9, 11))
+
+ // Passes typed columns to untyped `Dataset.select` API.
+ val err = intercept[AnalysisException] {
+ df.select(fooAgg(1), fooAgg(2), fooAgg(3), fooAgg(4), fooAgg(5),
fooAgg(6))
Review comment:
Yea, to be clear, if we add a 6th overload of typed `select`, a call to the
untyped `select` with 6 typed `count` could return `Dataset[(Long, Long, ...)]`
instead of `DataFrame`.
I think you meant something like existing `selectUntyped`? Although its
naming is confusing.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
