| From: | Shridhar Daithankar <shridhar_daithankar(at)persistent(dot)co(dot)in> | 
|---|---|
| To: | Adrian Demaestri <ademaestri(at)yahoo(dot)com> | 
| Cc: | pgsql-performance(at)postgresql(dot)org | 
| Subject: | Re: planner doesn't use multicolumn index | 
| Date: | 2003-10-08 14:27:38 | 
| Message-ID: | 3F841EDA.7000207@persistent.co.in | 
| Views: | Whole Thread | Raw Message | Download mbox | Resend email | 
| Thread: | |
| Lists: | pgsql-performance | 
Adrian Demaestri wrote:
> We've a table with about 8 million rows, and we need to get rows by the value 
 >of two of its fields( the type of the fields are int2 and int4,
>the where condition is v.g. partido=99 and partida=123). We created a
 >multicolumn index on that fields but the planner doesn't use it, it still use
 >a seqscan. That fields are primary key of the table and we clusterded the table
 >based on that index, but it still doesn't work. We also set the enviroment
 > variable enable_seqscan to false and nathing happends. The only way the
 >planner use it is in querys that order by the expression of the index.
Use partido=99::int2 and partida=123::int4
Match the data types basically..
Shridhar
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Manfred Koizar | 2003-10-08 14:35:24 | Re: planner doesn't use multicolumn index | 
| Previous Message | Tomasz Myrta | 2003-10-08 14:18:23 | Re: planner doesn't use multicolumn index |