Temp tables performance question

From: Hannu Krosing <hannu(at)tm(dot)ee>
To: PostgreSQL HACKERS <pgsql-hackers(at)postgresql(dot)org>
Subject: Temp tables performance question
Date: 2000-07-13 10:30:22
Message-ID: 396D9A3E.719489CE@tm.ee
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

I have a question about performance issues related to temporary tables.

IIRC temporary tables were implemented as ordinary tables with some
pre/post-processing to give them unique names so that they would not
clash with similar-named tables from other sessions.

Is this all that is done or has some work been done to improve their
performance further?

I'm mainly interested in INSERT performance as this is the area that is
much slower than other operations.

IMHO temporary tables could be made significantly faster than "ordinary"
as they are only accessed inside one session and thus have no need for
locking or even WAL as they could do with max 2 copies of the same row
the other of which can be discarded at end of transaction thereby making
it possible to provide much faster insert behaviour.

---------------------
Hannu

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Hannu Krosing 2000-07-13 11:50:22 Re: AW: Temp tables performance question
Previous Message Tatsuo Ishii 2000-07-13 10:21:16 Re: A postgreSQL question