Date: Sat, 19 May 2018 10:57:49 -0500
From: Justin Pryzby <pryzby@telsasoft.com>
To: "Kotapati, Anil" <Anil_Kotapati@comcast.com>
Cc: "pgsql-performance@postgresql.org" <pgsql-performance@postgresql.org>,
	"Kossi, Edem" <Edem_Kossi@comcast.com>,
	"Yadav, Rahul" <Rahul_Yadav2@comcast.com>,
	"Mukherjee, Pallavi (Contractor)" <Pallavi_Mukherjee@comcast.com>
Subject: Re: Help with tuning slow query
Message-ID: <20180519155749.GB30060@telsasoft.com>
References: <0AA18957-DDD4-413C-AE3A-2D3FD5EF0028@comcast.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <0AA18957-DDD4-413C-AE3A-2D3FD5EF0028@comcast.com>
User-Agent: Mutt/1.5.23 (2014-03-12)
Precedence: bulk

On Fri, May 18, 2018 at 08:32:55PM +0000, Kotapati, Anil wrote:
> We are facing issues with one of our query, when we use order by count it is taking lot of time to execute the query. To be precise it is taking 9 min to execute the query from table which has ~220 million records. Is there a way to make this query run faster and efficiently using order by count. Below is the query which I’m trying to run
> 
> Select account_number, sum(count_of_event) as "error_count"
> FROM event_daily_summary
> group by account_number,event_date,process_name
> having event_date >= '2018-05-07'
> and process_name='exp90d_xreerror'
> order by sum(count_of_event) desc
> limit 5000

Would you provide the information listed here ?  Table definition, query plan, etc
https://wiki.postgresql.org/wiki/Slow_Query_Questions

Also, why "HAVING" ? Shouldn't you use WHERE ?

Does the real query have conditions on event_date and process name or is that
just for testing purposes?

Justin