[SPARK-9029] [SQL] shortcut CaseKeyWhen if key is null

Author: Wenchen Fan <[email protected]>

Closes apache#7389 from cloud-fan/case-when and squashes the following commits:

ea4b6ba [Wenchen Fan] shortcut for case key when

Author: cloud-fan

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/conditionals.scala

@@ -230,24 +230,31 @@ case class CaseKeyWhen(key: Expression, branches: Seq[Expression]) extends CaseW
     }
   }
 
+  private def evalElse(input: InternalRow): Any = {
+    if (branchesArr.length % 2 == 0) {
+      null
+    } else {
+      branchesArr(branchesArr.length - 1).eval(input)
+    }
+  }
+
   /** Written in imperative fashion for performance considerations. */
   override def eval(input: InternalRow): Any = {
     val evaluatedKey = key.eval(input)
-    val len = branchesArr.length
-    var i = 0
-    // If all branches fail and an elseVal is not provided, the whole statement
-    // defaults to null, according to Hive's semantics.
-    while (i < len - 1) {
-      if (threeValueEquals(evaluatedKey, branchesArr(i).eval(input))) {
-        return branchesArr(i + 1).eval(input)
+    // If key is null, we can just return the else part or null if there is no else.
+    // If key is not null but doesn't match any when part, we need to return
+    // the else part or null if there is no else, according to Hive's semantics.
+    if (evaluatedKey != null) {
+      val len = branchesArr.length
+      var i = 0
+      while (i < len - 1) {
+        if (evaluatedKey ==  branchesArr(i).eval(input)) {
+          return branchesArr(i + 1).eval(input)
+        }
+        i += 2
       }
-      i += 2
     }
-    var res: Any = null
-    if (i == len - 1) {
-      res = branchesArr(i).eval(input)
-    }
-    return res
+    evalElse(input)
   }
 
   override def genCode(ctx: CodeGenContext, ev: GeneratedExpressionCode): String = {
@@ -261,8 +268,7 @@ case class CaseKeyWhen(key: Expression, branches: Seq[Expression]) extends CaseW
       s"""
         if (!$got) {
           ${cond.code}
-          if (!${keyEval.isNull} && !${cond.isNull}
-             && ${ctx.genEqual(key.dataType, keyEval.primitive, cond.primitive)}) {
+          if (!${cond.isNull} && ${ctx.genEqual(key.dataType, keyEval.primitive, cond.primitive)}) {
             $got = true;
             ${res.code}
             ${ev.isNull} = ${res.isNull};
@@ -290,19 +296,13 @@ case class CaseKeyWhen(key: Expression, branches: Seq[Expression]) extends CaseW
       boolean ${ev.isNull} = true;
       ${ctx.javaType(dataType)} ${ev.primitive} = ${ctx.defaultValue(dataType)};
       ${keyEval.code}
-      $cases
+      if (!${keyEval.isNull}) {
+        $cases
+      }
       $other
     """
   }
 
-  private def threeValueEquals(l: Any, r: Any) = {
-    if (l == null || r == null) {
-      false
-    } else {
-      l == r
-    }
-  }
-
   override def toString: String = {
     s"CASE $key" + branches.sliding(2, 2).map {
       case Seq(cond, value) => s" WHEN $cond THEN $value"