Re: Why Select Count(*) from table - took over 20 minutes?

On Tue, Oct 26, 2010 at 4:30 PM, Diego Schulz <dschulz(at)gmail(dot)com> wrote:
> On Tue, Oct 26, 2010 at 2:18 PM, Ozz Nixon <ozznixon(at)gmail(dot)com> wrote:
>> I am the only user on this system right now, and one table select count(*) took over 20 minutes:
>>
>> wikitags exists and has 58,988,656 records.
>>
>> Structure (in pascal) is:
>>
>>   quer.SQL.Add('create table '+DBTags+' (');
>>   quer.SQL.Add('   pagename '+SQL_TITLE+'(100) not null,');
>>   quer.SQL.Add('   tagword '+SQL_TITLE+'(15) not null,');
>>   quer.SQL.Add('   soundex2 '+SQL_TITLE+'(4) not null,');
>>   quer.SQL.Add('   metaphone '+SQL_TITLE+'(15) not null,');
>>   quer.SQL.Add('   metaphone2 '+SQL_TITLE+'(22) not null,');
>>   quer.SQL.Add('   carverphone '+SQL_TITLE+'(22) not null,');
>>   quer.SQL.Add('   instances '+SQL_INT32+' not null,');
>>   if SQL_NAME_PRIMARY_KEYS then quer.SQL.Add('   constraint '+DBTags+'_PK');
>>   quer.SQL.Add('   primary key(pagename, tagword, instances)');
>>   quer.SQL.Add(')');
>>
>> where SQL_TITLE = 'varchar', SQL_IN32 = 'int'
>>
>> I have hung off indexes for each column, to resolve my previous "performance" issue from 3+ weeks ago. However, COUNT() is still dog slow - this table is a write once, read many... *never* update, nor delete.
>>
>> Any suggestions?
>> --
>> Sent via pgsql-general mailing list (pgsql-general(at)postgresql(dot)org)
>> To make changes to your subscription:
>> http://www.postgresql.org/mailpref/pgsql-general
>>
>
> You could try something like what's suggested in this blog post:
> http://jakub.fedyczak.net/post/26
>
> I didn't actually tried it, but I think it should work ok.

Before you try that, you should ask yourself if you really need a 100%
accurate count. A reasonable approximation is maintained via the
stats system (pg_class.reltuples) that will often do and is free.

merlin