pg.ddx.io  pgsql-sql@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Mitch Vincent <mitch@venux.net>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-sql@postgresql.org
Subject: Re: Timestamp indexes
Date: Fri, 21 Jul 2000 13:40:19 -0400
Message-ID: <004901bff33a$bd41ebd0$4100000a@doot> (raw)
References: <001d01bff323$39cdb250$4100000a@doot>
	<13354.964192504@sss.pgh.pa.us>
	<007701bff32b$ce21d410$4100000a@doot>
	<13585.964200369@sss.pgh.pa.us>

With enable_seqscan off (Same query)

Sort  (cost=9282.89..9282.89 rows=4880 width=611)
  ->  Index Scan using applicants_created, applicants_resubmitted on
applicants a  (cost=0.00..8983.92 rows=4880 width=611)

...and..

! system usage stats:
!       7.541906 elapsed 5.368217 user 2.062897 system sec
!       [5.391668 user 2.070713 sys total]
!       1/543 [2/543] filesystem blocks in/out
!       0/9372 [0/9585] page faults/reclaims, 0 [0] swaps
!       0 [0] signals rcvd, 0/3 [4/7] messages rcvd/sent
!       7/101 [12/107] voluntary/involuntary context switches
! postgres usage stats:
!       Shared blocks:          0 read,          0 written, buffer hit rate
= 100.00%
!       Local  blocks:          0 read,          0 written, buffer hit rate
= 0.00%
!       Direct blocks:          0 read,          0 written
CommitTransactionCommand

Looks like that index scan is very unattractive... I'll look for some other
ways to speed up the query a bit..

Thanks!

-Mitch

----- Original Message -----
From: "Tom Lane" <tgl@sss.pgh.pa.us>
To: "Mitch Vincent" <mitch@venux.net>
Cc: <pgsql-sql@postgresql.org>
Sent: Friday, July 21, 2000 1:26 PM
Subject: Re: [SQL] Timestamp indexes


> "Mitch Vincent" <mitch@venux.net> writes:
> > select * from applicants as a where (a.created::date > '05-01-2000' or
> > a.resubmitted::date > '05-01-2000') order by (case when a.resubmitted >
> > a.created then a.resubmitted else a.created end) desc limit 10 offset 0
>
> > There is one of the queries.. I just remembered that the order by was
added
> > since last time I checked it's PLAN (in the 6.5.X days) -- could that be
the
> > problem?
>
> Probably.  With the ORDER BY in there, the LIMIT no longer applies
> directly to the scan (since a separate sort step is going to be
> necessary).  Now it's looking at a lot more data to be fetched by
> the scan, not just 10 records, so the indexscan becomes less attractive.
>
> Might be interesting to compare the estimated and actual runtimes
> between this query and what you get with "set enable_seqscan to off;"
>
> regards, tom lane
>




view thread (6+ messages)  latest in thread

Message-ID: <004901bff33a$bd41ebd0$4100000a@doot>
Permalink:  ../004901bff33a$bd41ebd0$4100000a@doot/
Also on:    postgresql.org/message-id/004901bff33a$bd41ebd0$4100000a@doot

 · 

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients from your mail client; the link fills in
  the headers below: reply via email
  Some clients, webmail especially, drop In-Reply-To and References from
  a link, and the reply then starts a new thread: use the mbox below.

  To: pgsql-sql@postgresql.org
  Cc: mitch@venux.net, tgl@sss.pgh.pa.us
  Subject: Re: Timestamp indexes
  In-Reply-To: <004901bff33a$bd41ebd0$4100000a@doot>
  References: <001d01bff323$39cdb250$4100000a@doot> <13354.964192504@sss.pgh.pa.us> <007701bff32b$ce21d410$4100000a@doot> <13585.964200369@sss.pgh.pa.us> <004901bff33a$bd41ebd0$4100000a@doot>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox