Re: Check lateral references within PHVs for memoize cache keys - Mailing list pgsql-hackers

From Richard Guo
Subject Re: Check lateral references within PHVs for memoize cache keys
Date
Msg-id CAMbWs49dEHRPe8poM_K39r2uOsaOZcg+Y0B5a8tF7vW3uVR3mw@mail.gmail.com
Whole thread Raw
In response to Re: Check lateral references within PHVs for memoize cache keys  (David Rowley <dgrowleyml@gmail.com>)
List pgsql-hackers

On Sun, Jul 9, 2023 at 12:17 PM David Rowley <dgrowleyml@gmail.com> wrote:
I just pushed a fix for this.  Thanks for reporting it.

BTW, I noticed a typo in the comment inside paraminfo_get_equal_hashops.

    foreach(lc, innerrel->lateral_vars)
    {
        Node       *expr = (Node *) lfirst(lc);
        TypeCacheEntry *typentry;

        /* Reject if there are any volatile functions in PHVs */
        if (contain_volatile_functions(expr))
        {
            list_free(*operators);
            list_free(*param_exprs);
            return false;
        }

The expressions in RelOptInfo.lateral_vars are not necessarily from
PHVs.  For instance

explain (costs off)
select * from t t1 join
    lateral (select * from t t2 where t1.a = t2.a offset 0) on true;
            QUERY PLAN
----------------------------------
 Nested Loop
   ->  Seq Scan on t t1
   ->  Memoize
         Cache Key: t1.a
         Cache Mode: binary
         ->  Seq Scan on t t2
               Filter: (t1.a = a)
(7 rows)

The lateral Var 't1.a' comes from the lateral subquery, not PHV.

This seems a typo from 63e4f13d.  How about we change it to the below?

-     /* Reject if there are any volatile functions in PHVs */
+     /* Reject if there are any volatile functions in lateral vars */

Thanks
Richard

pgsql-hackers by date:

Previous
From: Alvaro Herrera
Date:
Subject: Re: Consistent coding for the naming of LR workers
Next
From: Tomas Vondra
Date:
Subject: Re: Configurable FP_LOCK_SLOTS_PER_BACKEND