kgyrtkirk commented on a change in pull request #544: HIVE-16924 Support 
distinct in presence of Group By
URL: https://github.com/apache/hive/pull/544#discussion_r259778186
 
 

 ##########
 File path: ql/src/java/org/apache/hadoop/hive/ql/parse/SemanticAnalyzer.java
 ##########
 @@ -4194,27 +4191,29 @@ public static long unsetBit(long bitmap, int bitIdx) {
   }
 
   /**
-   * This function is a wrapper of parseInfo.getGroupByForClause which
-   * automatically translates SELECT DISTINCT a,b,c to SELECT a,b,c GROUP BY
-   * a,b,c.
+   * Returns the GBY, if present;
+   * DISTINCT, if present, will be handled when generating the SELECT.
    */
   List<ASTNode> getGroupByForClause(QBParseInfo parseInfo, String dest) throws 
SemanticException {
-    if (parseInfo.getSelForClause(dest).getToken().getType() == 
HiveParser.TOK_SELECTDI) {
-      ASTNode selectExprs = parseInfo.getSelForClause(dest);
-      List<ASTNode> result = new ArrayList<ASTNode>(selectExprs == null ? 0
-          : selectExprs.getChildCount());
-      if (selectExprs != null) {
-        for (int i = 0; i < selectExprs.getChildCount(); ++i) {
-          if (((ASTNode) selectExprs.getChild(i)).getToken().getType() == 
HiveParser.QUERY_HINT) {
+    // When *not* invoked by CalcitePlanner, return the DISTINCT as a GBY
+    // CBO will handle the DISTINCT in 
CalcitePlannerAction.genSelectLogicalPlan
+    ASTNode selectExpr = parseInfo.getSelForClause(dest);
+    Collection<ASTNode> aggregateFunction = 
parseInfo.getDestToAggregationExprs().get(dest).values();
+    if (isSelectDistinct(selectExpr) && !isGroupBy(selectExpr) && 
!isAggregateInSelect(selectExpr, aggregateFunction)) {
 
 Review comment:
   what plan are we ending up if we have select distinct which has an aggregate 
(this condition will be false) and the cbo is not working? is there a test for 
that?
   

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on 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

Reply via email to