Actually, this issue has been fixed by 3684fd2.

apache · Jan 12, 2015 · 6cfadd2 · 6cfadd2
1 parent d42b707
commit 6cfadd2
Show file tree

Hide file tree

Showing 4 changed files with 19 additions and 17 deletions.
diff --git a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala
@@ -209,10 +209,10 @@ class Analyzer(catalog: Catalog,
 
           aggregateExprs.find { e =>
             !isValidAggregateExpression(e.transform {
-              // Should trim aliases. These aliases can be introduced while resolving unnamed
-              // expressions like `GetField` and UDF calls, and when a user explicitly assigns
-              // an alias to a grouping expression in the SELECT clause.
-              case a: Alias => a.child
+              // Should trim aliases around `GetField`s. These aliases are introduced while
+              // resolving struct field accesses, because `GetField` is not a `NamedExpression`.
+              // (Should we just turn `GetField` into a `NamedExpression`?)
+              case Alias(g: GetField, _) => g
             })
           }.foreach { e =>
             throw new TreeNodeException(plan, s"Expression not in GROUP BY: $e")

diff --git a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/planning/patterns.scala b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/planning/patterns.scala
@@ -143,8 +143,6 @@ object PartialAggregation {
         // referenced in the second aggregation.
         val namedGroupingExpressions: Map[Expression, NamedExpression] = groupingExpressions.map {
           case n: NamedExpression => (n, n)
-          // TODO: When a UDF is used in the group by clause, for example, GROUP BY YEAR(...),
-          // PartialGroup will not be a proper name.
           case other => (other, Alias(other, "PartialGroup")())
         }.toMap
 

diff --git a/sql/core/src/test/scala/org/apache/spark/sql/SQLQuerySuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/SQLQuerySuite.scala
@@ -997,17 +997,6 @@ class SQLQuerySuite extends QueryTest with BeforeAndAfterAll {
     dropTempTable("data")
   }
 
-  test("SPARK-4296 Grouping field with UDF as sub expression") {
-    registerFunction("triple", (_: Int) * 3)
-    jsonRDD(sparkContext.makeRDD("""{"a": 1}""" :: Nil)).registerTempTable("data")
-    checkAnswer(sql("SELECT triple(a) FROM data GROUP BY triple(a)"), 3)
-    dropTempTable("data")
-
-    jsonRDD(sparkContext.makeRDD("""{"a": 1}""" :: Nil)).registerTempTable("data")
-    checkAnswer(sql("SELECT triple(a) + 1 FROM data GROUP BY triple(a) + 1"), 4)
-    dropTempTable("data")
-  }
-
   test("SPARK-4432 Fix attribute reference resolution error when using ORDER BY") {
     checkAnswer(
       sql("SELECT a + b FROM testData2 ORDER BY a"),

diff --git a/sql/hive/src/test/scala/org/apache/spark/sql/hive/execution/SQLQuerySuite.scala b/sql/hive/src/test/scala/org/apache/spark/sql/hive/execution/SQLQuerySuite.scala
@@ -214,4 +214,19 @@ class SQLQuerySuite extends QueryTest {
         Seq.empty[Row])
     }
   }
+
+  test("SPARK-4296 Grouping field with Hive UDF as sub expression") {
+    val rdd = sparkContext.makeRDD("""{"a": "str", "b":"1", "c":"1970-01-01 00:00:00"}""" :: Nil)
+    jsonRDD(rdd).registerTempTable("data")
+    checkAnswer(
+      sql("SELECT concat(a, '-', b), year(c) FROM data GROUP BY concat(a, '-', b), year(c)"),
+      Seq(Seq("str-1", 1970)))
+
+    dropTempTable("data")
+
+    jsonRDD(rdd).registerTempTable("data")
+    checkAnswer(sql("SELECT year(c) + 1 FROM data GROUP BY year(c) + 1"), Seq(Seq(1971)))
+
+    dropTempTable("data")
+  }
 }