MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII;
 format=flowed
Content-Transfer-Encoding: 7bit
Date: Thu, 14 Sep 2017 14:03:20 +0200
From: vinny <vinny@xs4all.nl>
To: Subramaniam C <subramaniam31784@gmail.com>
Cc: pgsql-performance@postgresql.org
Subject: Re: Store/Retrieve time series data from PostgreSQL
In-Reply-To: 
 <CAL=06WmUGXv2Td30XaT9mHYSWJCVombq85+z2Th+0UPs7sOKsQ@mail.gmail.com>
References: 
 <CAL=06WmUGXv2Td30XaT9mHYSWJCVombq85+z2Th+0UPs7sOKsQ@mail.gmail.com>
Message-ID: <1eb896cf2a6bdc3fd7e1567c86ecafe9@xs4all.nl>
User-Agent: XS4ALL Webmail
Precedence: bulk
Sender: pgsql-performance-owner@postgresql.org

On 2017-09-14 13:51, Subramaniam C wrote:
> Hi
> 
> QUERY :-
> 
> _select distinct on (health_timeseries.mobid) mobid,
> health_timeseries.health, health_timeseries.hour from
> health_timeseries where hour >=(1505211054000/(3600*1000))-1 and hour
> <= 1505211054000/(3600*1000) ORDER BY health_timeseries.mobid DESC,
> health_timeseries.hour DESC;_
> 

Did you run EXPLAIN on this query to see what it is actually doing?

What you are doing how is selecting all rows from the last hour,
sorting them by mobid and hour, and then DISTINCT filters out al 
duplicates.

Sorting on mobid is therefor useless, DISTINCT still has to check all 
rows.

Sorting on mobid and hour will take a long time if there is no index for 
it,
so if you don't have an index on the mobid and hour together then you 
should probably try that.


But, see what EXPLAIN tells you first.

Regards,
Vincent.


-- 
Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance