Re: Unexpected (wrong?) result querying boolean partitioned table with NULL partition

From: Richard Guo <guofenglinux(at)gmail(dot)com>
To: David Rowley <dgrowleyml(at)gmail(dot)com>
Cc: David Kimura <david(dot)g(dot)kimura(at)gmail(dot)com>, PostgreSQL Development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Unexpected (wrong?) result querying boolean partitioned table with NULL partition
Date: 2023-04-13 02:39:03
Message-ID: CAMbWs48x1jvE8Ab67Xbe8odLh96uLhw=DQi6oiQqBTvkyD7xnw@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Wed, Apr 12, 2023 at 7:13 PM David Rowley <dgrowleyml(at)gmail(dot)com> wrote:

> There's already code to effectively handle <> operators. Just the
> PartClauseInfo.op_is_ne needs to be set to true.
> get_matching_list_bounds() then handles that by taking the inverse of
> the partitions matching the equality operator.
>
> Effectively, I think that's the attached patch.

I think there is a thinko here.

+ switch (btest->booltesttype)
+ {
+ case IS_NOT_TRUE:
+ *noteq = true;
+ /* fall through */
+ case IS_TRUE:
+ *outconst = (Expr *) makeBoolConst(true, false);
+ break;
+ case IS_NOT_FALSE:
+ *noteq = true;
+ /* fall through */
+ case IS_FALSE:
+ *outconst = (Expr *) makeBoolConst(false, false);
+ break;
+ default:
+ Assert(false); /* hmm? */
+ return PARTCLAUSE_UNSUPPORTED;
+ }

The *outconst should be set to true in case IS_NOT_FALSE and set to
false in case IS_NOT_TRUE, something like:

switch (btest->booltesttype)
{
- case IS_NOT_TRUE:
+ case IS_NOT_FALSE:
*noteq = true;
/* fall through */
case IS_TRUE:
*outconst = (Expr *) makeBoolConst(true, false);
break;
- case IS_NOT_FALSE:
+ case IS_NOT_TRUE:
*noteq = true;
/* fall through */
case IS_FALSE:

Thanks
Richard

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Smith 2023-04-13 02:42:05 Re: pg_upgrade and logical replication
Previous Message Amit Kapila 2023-04-13 02:32:07 Re: Non-superuser subscription owners