>> Hm. I'd sort of expect the "z" to become both the table and column
authorBruce Momjian
Tue, 6 Aug 2002 05:34:10 +0000 (05:34 +0000)
committerBruce Momjian
Tue, 6 Aug 2002 05:34:10 +0000 (05:34 +0000)
>> alias in this case.  What do you think?
>
> I guess that would make sense. I'll make a separate patch just for
that
> change if that's OK.
>

Simple change -- patch attached.

test=# select * from myfoo1() as z;
  z
----
   1
   2
   3
(3 rows)

Joe Conway

src/backend/parser/parse_relation.c

index 6732dc54b5331a5ed2100c605e4a1b21f12fc776..8a15c0f7cbff4f52a9504a5b998e26f2d96875a4 100644 (file)
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *   $Header: /cvsroot/pgsql/src/backend/parser/parse_relation.c,v 1.74 2002/08/06 05:33:29 momjian Exp $
+ *   $Header: /cvsroot/pgsql/src/backend/parser/parse_relation.c,v 1.75 2002/08/06 05:34:10 momjian Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -807,7 +807,7 @@ addRangeTableEntryForFunction(ParseState *pstate,
            elog(ERROR, "Too many column aliases specified for function %s",
                 funcname);
        if (numaliases == 0)
-           eref->colnames = makeList1(makeString(funcname));
+           eref->colnames = makeList1(makeString(eref->aliasname));
    }
    else if (functyptype == 'p' && funcrettype == RECORDOID)
    {