Re: Select in FOR LOOP Performance

From: Tom Lane
Subject: Re: Select in FOR LOOP Performance
Date: ,
(view: Whole thread, Raw)
In response to: Select in FOR LOOP Performance  (Charles Joseph)
List: pgsql-performance

Charles Joseph <> writes:
>     I face problem when running the following pgplsql
> function. The problem is it takes more than 24hours to
> complete
>     the calculation. The EMP table has about 200,000
> records.

Sure there are no infinite loops of PARENT links in your table?

Also, if CODE is supposed to be unique, you should probably declare
its index that way.  Or at least make sure the planner knows it's
unique (have you ANALYZEd the table lately?)

            regards, tom lane

pgsql-performance by date:

From: Alex Turner
Subject: Re: multi billion row tables: possible or insane?
From: Tom Lane
Subject: Re: MAIN vs. PLAIN