From: | Hans-Juergen Schoenig <postgres(at)cybertec(dot)at> |
---|---|
To: | Peter Eisentraut <peter_e(at)gmx(dot)net> |
Cc: | pgsql-hackers(at)postgresql(dot)org, Justin <justin(at)emproshunts(dot)com> |
Subject: | Re: Ad Hoc Indexes |
Date: | 2008-02-18 20:16:38 |
Message-ID: | 16AB0F44-B51C-4CB7-9748-EBFDB416847A@cybertec.at |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Feb 18, 2008, at 9:09 PM, Peter Eisentraut wrote:
> Justin wrote:
>> Now for my question It does not appear PostgreSQL does not have an Ad
>> Hoc Indexes ability where the Query planner will create an in memory
>> index based on the Select, Update, Insert or Delete commands.
>
> How is that supposed to work? In order to create an index you
> would need to
> visit all the rows in the table. If you do that, you could just as
> well
> answer the query off a sequential scan.
>
this is not quite true.
this kind of indexing makes sense if you visit the same data over and
over again. WITH-queries would be an example for that and self joins
could benefit from the this feature too.
the question however is: why not create "normal indexes" straight away?
i am not sure if the benefit of ad-hoc indexes justify additional
complexity in the code ...
best regards,
hans
--
Cybertec Schönig & Schönig GmbH
PostgreSQL Solutions and Support
Gröhrmühlgasse 26, 2700 Wiener Neustadt
Tel: +43/1/205 10 35 / 340
www.postgresql.at, www.cybertec.at
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2008-02-18 20:18:17 | Re: Ad Hoc Indexes |
Previous Message | Tom Lane | 2008-02-18 20:13:15 | Re: IDENTITY/GENERATED patch |