An oversight in ExecInitAgg for grouping sets - Mailing list pgsql-hackers

From Richard Guo
Subject An oversight in ExecInitAgg for grouping sets
Date
Msg-id CAMbWs4-vZuuPOZsKOYnSAaPYGKhmacxhki+vpOKk0O7rymccXQ@mail.gmail.com
Whole thread Raw
Responses Re: An oversight in ExecInitAgg for grouping sets
List pgsql-hackers
I happened to notice $subject.  It happens when we build eqfunctions for
each grouping set.

     /* for each grouping set */
     for (int k = 0; k < phasedata->numsets; k++)
     {
         int         length = phasedata->gset_lengths[k];

         if (phasedata->eqfunctions[length - 1] != NULL)
             continue;

         phasedata->eqfunctions[length - 1] =
             execTuplesMatchPrepare(scanDesc,
                                    length,
                                    aggnode->grpColIdx,
                                    aggnode->grpOperators,
                                    aggnode->grpCollations,
                                    (PlanState *) aggstate);
     }

If it is an empty grouping set, its length will be zero, and accessing
phasedata->eqfunctions[length - 1] is not right.

I think we can just skip building the eqfunctions for empty grouping
set.

--- a/src/backend/executor/nodeAgg.c
+++ b/src/backend/executor/nodeAgg.c
@@ -3494,6 +3494,10 @@ ExecInitAgg(Agg *node, EState *estate, int eflags)
                {
                    int         length = phasedata->gset_lengths[k];

+                   /* skip empty grouping set */
+                   if (length == 0)
+                       continue;
+
                    if (phasedata->eqfunctions[length - 1] != NULL)
                        continue;

Thanks
Richard

pgsql-hackers by date:

Previous
From: "Takamichi Osumi (Fujitsu)"
Date:
Subject: RE: Time delayed LR (WAS Re: logical replication restrictions)
Next
From: Masahiko Sawada
Date:
Subject: Re: Perform streaming logical transactions by background workers and parallel apply