>I have a very large table (~5GB, 100mm rows) on which I am creating an
>index on an int4 column:
>CREATE INDEX CONCURRENTLY rums_idx2
> ON rums
> USING hash
> (user_id);
Why are you using a hash index on an int4 column?
as you have 100mm (mm= Million?) rows, only 1 GIG ram and the
column_name is "user_id", I suspect:
- you have far less then 100*10E6 Users
- there are less distince user_ids than rows
I just know "hash" from general programming, thatfor my belly is
announcing "hash collisions, hash collisions" ...
(within PostgreSQL I once read that hash indices are only for some
cases I could not match to something that happens in my world; since
then I am using default btree and am VERY happy.)
Best wishes,
Harald
--
GHUM Harald Massa
persuadere et programmare
Harald Armin Massa
Spielberger Straße 49
70435 Stuttgart
0173/9409607
fx 01212-5-13695179
-
EuroPython 2008 will take place in Vilnius, Lithuania - Stay tuned!