| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | "Pavel Stehule" <pavel(dot)stehule(at)gmail(dot)com> |
| Cc: | "PostgreSQL-development Hackers" <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: possible minor EXPLAIN bug? |
| Date: | 2008-08-19 14:25:02 |
| Message-ID: | 20052.1219155902@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
"Pavel Stehule" <pavel(dot)stehule(at)gmail(dot)com> writes:
> I thing so Agg node doesn't set width well:
The planner doesn't really bother to set the width correctly for any
expression-computing node. This is partly laziness, but OTOH it is
very hard to estimate a sane width for any function returning a
variable-width data type; eg what are the odds of a useful answer for
select repeat(textcol, intcol) from tab1;
For plan nodes that return just Vars it's easier, since we generally
have got stats about average column widths.
I think Agg just copies the width of its input ...
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2008-08-19 14:40:53 | Re: Compatibility types, type aliases, and distinct types |
| Previous Message | Markus Wanner | 2008-08-19 12:04:46 | Re: Postgres-R |