[SPARK-12727][SQL] support SQL generation for aggregate with multi-distinct

## What changes were proposed in this pull request?

This PR add SQL generation support for aggregate with multi-distinct, by simply moving the `DistinctAggregationRewriter` rule to optimizer.

More discussions are needed as this breaks an import contract: analyzed plan should be able to run without optimization.  However, the `ComputeCurrentTime` rule has kind of broken it already, and I think maybe we should add a new phase for this kind of rules, because strictly speaking they don't belong to analysis and is coupled with the physical plan implementation.

## How was this patch tested?

existing tests

Author: Wenchen Fan <wenchen@databricks.com>

Closes #11579 from cloud-fan/distinct.

1 files changed, 1 insertions, 4 deletions

diff --git a/sql/hive/src/test/scala/org/apache/spark/sql/hive/LogicalPlanToSQLSuite.scala b/sql/hive/src/test/scala/org/apache/spark/sql/hive/LogicalPlanToSQLSuite.scala
index ed85856f01..ccce9871e2 100644
--- a/sql/hive/src/test/scala/org/apache/spark/sql/hive/LogicalPlanToSQLSuite.scala
+++ b/sql/hive/src/test/scala/org/apache/spark/sql/hive/LogicalPlanToSQLSuite.scala
@@ -139,7 +139,6 @@ class LogicalPlanToSQLSuite extends SQLBuilderTest with SQLTestUtils {
       """.stripMargin)
   }
 
-
   test("intersect") {
     checkHiveQl("SELECT * FROM t0 INTERSECT SELECT * FROM t0")
   }
@@ -367,9 +366,7 @@ class LogicalPlanToSQLSuite extends SQLBuilderTest with SQLTestUtils {
     checkHiveQl("SELECT * FROM parquet_t0 TABLESAMPLE(0.1 PERCENT) WHERE 1=0")
   }
 
-  // TODO Enable this
-  // Query plans transformed by DistinctAggregationRewriter are not recognized yet
-  ignore("multi-distinct columns") {
+  test("multi-distinct columns") {
     checkHiveQl("SELECT a, COUNT(DISTINCT b), COUNT(DISTINCT c), SUM(d) FROM parquet_t2 GROUP BY a")
   }