Re: PostgreSQL OR performance - Mailing list pgsql-performance

From Richard Huxton
Subject Re: PostgreSQL OR performance
Date
Msg-id 49131ED5.80307@archonet.com
Whole thread Raw
In response to Re: PostgreSQL OR performance  ("Віталій Тимчишин" <tivv00@gmail.com>)
Responses Re: PostgreSQL OR performance  ("Віталій Тимчишин" <tivv00@gmail.com>)
List pgsql-performance
Віталій Тимчишин wrote:
> As you can see from other plans, it do have all the indexes to perform it's
> work fast (when given part by part). It simply do not wish to use them. My
> question: Is this a configuration problem or postgresql optimizer simply
> can't do such a query rewrite?

I must admit, I haven't managed to figure out what your query is trying
to do, but then that's a common problem with autogenerated queries.

The main question that needs answering is why the planner thinks you're
going to get 1.3 billion rows in the "or" query:

"Nested Loop  (cost=4588.13..960900482668.95 rows=1386158171 width=32)"

You don't show "explain analyse" for this query, so there's no way of
knowing how many rows get returned but presumably you're expecting
around 88000. What does "explain analyse" return?

--
  Richard Huxton
  Archonet Ltd

pgsql-performance by date:

Previous
From: "Віталій Тимчишин"
Date:
Subject: Re: PostgreSQL OR performance
Next
From: "Helio Campos Mello de Andrade"
Date:
Subject: Re: PostgreSQL OR performance