| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Viktor Rosenfeld <rosenfel(at)informatik(dot)hu-berlin(dot)de> |
| Cc: | pgsql-general(at)postgresql(dot)org |
| Subject: | Re: UPDATE runs slow in a transaction |
| Date: | 2008-07-14 18:52:27 |
| Message-ID: | 29834.1216061547@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
Viktor Rosenfeld <rosenfel(at)informatik(dot)hu-berlin(dot)de> writes:
> the script below runs very fast when executed alone. But when I call
> it from within a transaction block it's so slow that I have to abort
> it after a while. Specifically the second-to-last UPDATE seems to
> take forever within a transaction while it completes in about 3
> seconds outside a transaction.
Since the table you're working on was just created in the same
transaction, there's been no opportunity for autovacuum to run an
ANALYZE on it; that's probably preventing selection of a good plan.
Try throwing in an "ANALYZE tmp" after you load the table.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2008-07-14 19:02:04 | Re: statement timeout and pg_dump |
| Previous Message | Tom Lane | 2008-07-14 18:20:14 | Re: How to remove the possibility to search the parts of a compound word with hyphen |