| From: | Hannu Krosing <hannu(at)skype(dot)net> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | Gregory Stark <stark(at)enterprisedb(dot)com>, Kenneth Marshall <ktm(at)rice(dot)edu>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Hash index todo list item |
| Date: | 2007-09-06 10:24:48 |
| Message-ID: | 1189074288.7470.5.camel@hannu-laptop |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Ühel kenal päeval, E, 2007-09-03 kell 19:55, kirjutas Tom Lane:
> Gregory Stark <stark(at)enterprisedb(dot)com> writes:
> > "Kenneth Marshall" <ktm(at)rice(dot)edu> writes:
> >> - What about multi-column indexes? The current implementation
> >> only supports 1 column.
>
> > That seems kind of weird. It seems obvious that you mix the three hashes
> > together which reduces it to the solved problem.
>
> No, because part of the deal is that you can do lookups using only the
> leading index columns. At least, all the existing multicolumn index
> types can do that.
One approahc is not to mix hashes, but to partition the hash, so that
each column gets its N bits in the hash.
If you do it smartly you can use any column for index lookups, not just
the leading one.
> regards, tom lane
>
> ---------------------------(end of broadcast)---------------------------
> TIP 9: In versions below 8.0, the planner will ignore your desire to
> choose an index scan if your joining column's datatypes do not
> match
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Richard Huxton | 2007-09-06 10:30:17 | Re: [FEATURE REQUEST] Streaming Onlinebackup (Maybe OFFTOPIC) |
| Previous Message | apoc9009 | 2007-09-06 10:21:10 | Re: [FEATURE REQUEST] Streaming Onlinebackup (Maybe OFFTOPIC) |