[SPARK-20409][SQL] fail early if aggregate function in GROUP BY

## What changes were proposed in this pull request?

It's illegal to have aggregate function in GROUP BY, and we should fail at analysis phase, if this happens.

## How was this patch tested?

new regression test

Author: Wenchen Fan <wenchen@databricks.com>

Closes #17704 from cloud-fan/minor.

1 files changed, 7 insertions, 0 deletions

diff --git a/sql/core/src/test/scala/org/apache/spark/sql/DataFrameAggregateSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/DataFrameAggregateSuite.scala
index e7079120bb..8569c2d76b 100644
--- a/sql/core/src/test/scala/org/apache/spark/sql/DataFrameAggregateSuite.scala
+++ b/sql/core/src/test/scala/org/apache/spark/sql/DataFrameAggregateSuite.scala
@@ -538,4 +538,11 @@ class DataFrameAggregateSuite extends QueryTest with SharedSQLContext {
       Seq(Row(3, 0, 0.0, 1, 5.0), Row(2, 1, 4.0, 0, 0.0))
     )
   }
+
+  test("aggregate function in GROUP BY") {
+    val e = intercept[AnalysisException] {
+      testData.groupBy(sum($"key")).count()
+    }
+    assert(e.message.contains("aggregate functions are not allowed in GROUP BY"))
+  }
 }