Yi Zhao wrote:
> yes, I thinks it is unrelated with index, beacause the problem is still
> exist after reindex.
>
> I try to change the work memory or shared memory, it's no use:(
Of course not.
Please post the schema of involved tables, and enough data in them to be
able to reproduce the problem.
--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.