Re: like performance w/o wildcards. - Mailing list pgsql-general

From Tom Lane
Subject Re: like performance w/o wildcards.
Date
Msg-id 1723.1059949684@sss.pgh.pa.us
Whole thread Raw
In response to like performance w/o wildcards.  (Joseph Shraibman <jks@selectacast.net>)
Responses Re: like performance w/o wildcards.  (Joseph Shraibman <jks@selectacast.net>)
List pgsql-general
Joseph Shraibman <jks@selectacast.net> writes:
> I notice in 7.3.3 the planner can't tell when a LIKE has no wildcards
> and is in reality an '='.  Is this an easy change to make?

On what do you base that conclusion?

regression=# create table t1 (f1 text unique);
NOTICE:  CREATE TABLE / UNIQUE will create implicit index 't1_f1_key' for table 't1'
CREATE TABLE
regression=# explain select * from t1 where f1 like 'foo';
                             QUERY PLAN
---------------------------------------------------------------------
 Index Scan using t1_f1_key on t1  (cost=0.00..4.82 rows=1 width=32)
   Index Cond: (f1 = 'foo'::text)
   Filter: (f1 ~~ 'foo'::text)
(3 rows)

regression=#

            regards, tom lane

pgsql-general by date:

Previous
From: "Vernon Smith"
Date:
Subject: Re: Inheritance & multiple-value fields
Next
From: Ron Johnson
Date:
Subject: Re: Monthly table partitioning for fast purges?