| From: | "Joshua D(dot) Drake" <jd(at)commandprompt(dot)com> |
|---|---|
| To: | pgsql-hackers(at)postgresql(dot)org |
| Cc: | "Mark Woodward" <pgsql(at)mohawksoft(dot)com>, "Martijn van Oosterhout" <kleptog(at)svana(dot)org>, "Zdenek Kotala" <zdenek(dot)kotala(at)sun(dot)com> |
| Subject: | Re: update/insert, |
| Date: | 2006-07-05 15:52:12 |
| Message-ID: | 200607050852.12454.jd@commandprompt.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
> > Which is faster will probably depends on what is more common in your DB:
> > row already exists or not. If you know that 99% of the time the row
> > will exist, the update will probably be faster because you'll only
> > execute one query 99% of the time.
>
> OK, but the point of the question is that constantly updating a single row
> steadily degrades performance, would delete/insery also do the same?
Yes. Delete still creates a dead row. There are programatic ways around this
but keeping a delete table that can be truncated at intervals.
Joshua D. Drake
>
> ---------------------------(end of broadcast)---------------------------
> TIP 5: don't forget to increase your free space map settings
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Greg Stark | 2006-07-05 16:00:05 | Scan Keys |
| Previous Message | mark | 2006-07-05 15:25:18 | Re: update/insert, |