On Wed, 2003-08-20 at 14:51, Tom Lane wrote: > Bruno Wolff III <[EMAIL PROTECTED]> writes: > > On Wed, Aug 20, 2003 at 13:44:59 -0500, > > Ron Johnson <[EMAIL PROTECTED]> wrote: > >> The GROUP BY does implicit sorting, so an ORDER BY on the exact same > >> column(s) as the GROUP BY is redundant. > > > That is an implementation detail, not a promise. With hashed aggregates > > in 7.4, you might find this isn't true. > > s/might/will/
>From 7.3.3, where the records were randomly inserted; note how GROUP BY acts like I described: test1=# select f, count(*) test1-# from t test1-# group by f; f | count ---+------- 1 | 3 2 | 5 4 | 4 (3 rows) The new 7.4 attitude is *really* good to know, because, otherwise, all our reports would break! -- ----------------------------------------------------------------- Ron Johnson, Jr. [EMAIL PROTECTED] Jefferson, LA USA "Fair is where you take your cows to be judged." Unknown ---------------------------(end of broadcast)--------------------------- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match