[GENERAL] Aggregate query on large tables - Mailing list pgsql-general

From Job
Subject [GENERAL] Aggregate query on large tables
Date
Msg-id 88EF58F000EC4B4684700C2AA3A73D7A0817DC96CB17@W2008DC01.ColliniConsulting.lan
Whole thread Raw
Responses Re: [GENERAL] Aggregate query on large tables  (Bill Moran <wmoran@potentialtech.com>)
Re: [GENERAL] Aggregate query on large tables  (Adrian Klaver <adrian.klaver@aklaver.com>)
List pgsql-general
Hi,

i have a table with about 400 millions of rows and i need to build some aggregate function for reporting.
I noticed that query performances are slowing down, even though indexes are present.

Query is simple (i make an example, my table is in italian language):
select a,sum(b) from table where a=x and c=y group by a

a is a varchar
b is an integer

x and y are two field i use for filter results.

I tried to create different indexes to try to speed up performance:
index1 (a)
index2 (c)
index3 (a,c)

I noticed, with query planner, that the mulfi-field index is not used.
Postgresql 9.6.1 still use scan without indexes.

I obtain significant improvements only if i create a materialized view with aggregated data.

But i would like to avoid - if possible - creating (and mantaining) the materialized view.

Than you!
/F

pgsql-general by date:

Previous
From: Nicolas Paris
Date:
Subject: Re: [GENERAL] TimeScaleDB -- Open Source Time Series DatabaseReleased (www.i-programmer.info);
Next
From: Bill Moran
Date:
Subject: Re: [GENERAL] Aggregate query on large tables