Message-ID: <5145A5C6.9000703@gmail.com>
Date: Sun, 17 Mar 2013 12:15:18 +0100
From: Surfing <onlinesurfing@gmail.com>
User-Agent: Mozilla/5.0 (Windows NT 6.2; WOW64;
	rv:17.0) Gecko/20130307 Thunderbird/17.0.4
MIME-Version: 1.0
To: pgsql-sql@postgresql.org
Subject: Efficiency Problem
Content-Type: multipart/alternative;
	boundary="------------060603080809060206020106"
Precedence: bulk
Sender: pgsql-sql-owner@postgresql.org

This is a multi-part message in MIME format.
--------------060603080809060206020106
Content-Type: text/plain; charset=ISO-8859-15; format=flowed
Content-Transfer-Encoding: 7bit

Hi all,
     I'm composing a query from a web application of type:

*SELECT * FROM table WHERE a_text_field LIKE replace_something 
('%**/a_given_string/**%');*

The function replace_something( ... ) is a stored procedure that 
replaces some particular characters with others.
The problem is that I noticed that this query is inefficient... and I 
think that the replace_something ( ... ) function is called for each row 
of the table.

This observation is motivated by the fact that it takes around 30 
seconds to execute on the table (of about 25,000 rows), whereas if I 
execute:
*SELECT * FROM table WHERE a_text_field LIKE 
'**/pre_processed_string/**//**';*

where/pre_processed_string///is the result of the application of 
replace_something ('%/a_given_string/%')  it just takes 164ms.

The execution of
*SELECT replace_something ('%**/a_given_string/**%')*
  takes only 14ms.

Summarizing,
- Replace function:     14ms
- SELECT query without replace function:     164ms
- SELECT query with replace function:      30.000ms

Morever, I cannot create a stored procedure that precalculate the 
/pre_processed_string /and executes the query, since I dinamically
compose other conditions in the WHERE clause.

Any suggestion?

Thank you.
//

--------------060603080809060206020106
Content-Type: text/html; charset=ISO-8859-15
Content-Transfer-Encoding: 8bit

<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=ISO-8859-15">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    Hi all,<br>
        I'm composing a query from a web application of type:<br>
    <br>
        <b>SELECT * FROM table WHERE a_text_field LIKE
      replace_something ('%</b><b><i>a_given_string</i></b><b>%');</b><br>
    <br>
    The function replace_something( ... ) is a stored procedure that
    replaces some particular characters with others.<br>
    The problem is that I noticed that this query is inefficient... and
    I think that the replace_something ( ... ) function is called for
    each row of the table.<br>
    <br>
    This observation is motivated by the fact that it takes around 30
    seconds to execute on the table (of about 25,000 rows), whereas if I
    execute:<br>
        <b>SELECT * FROM table WHERE a_text_field LIKE '</b><b><i>pre_processed_string</i></b><b><i></i></b><b>';</b><br>
    <br>
    where<i> pre_processed_string</i><i> </i>is the result of the
    application of replace_something ('%<i>a_given_string</i>%')  it
    just takes 164ms.<br>
    <br>
    The execution of <br>
        <b>SELECT replace_something ('%</b><b><i>a_given_string</i></b><b>%')</b><br>
     takes only 14ms.<br>
    <br>
    Summarizing, <br>
    - Replace function:     14ms<br>
    - SELECT query without replace function:     164ms<br>
    - SELECT query with replace function:      30.000ms<br>
    <br>
    Morever, I cannot create a stored procedure that precalculate the <i>pre_processed_string
    </i>and executes the query, since I dinamically <br>
    compose other conditions in the WHERE clause.<br>
    <br>
    Any suggestion?<br>
    <br>
    Thank you.<br>
    <i></i>
  </body>
</html>

--------------060603080809060206020106--