Re: BUG #4913: Row missing from primary key index

Andrew Gierth wrote:
> >>>>> "Mathieu" == "Mathieu De Zutter" <mathieu@dezutter.org> writes:
>
>  Mathieu> I have a table log_event with a primary key on an integer
>  Mathieu> 'id', called log_event_pkey.
>
>  Mathieu> The tables contains a duplicate for id = 15723018. The
>  Mathieu> duplicate (note that besides the id, all data differs)
>  Mathieu> doesn't seem to be known by the index at all.
> [snip]
>
> This was first reported on IRC and I spent a little time working with
> the OP trying to dig up info to suggest a cause; here is the relevent
> data (all provided by the OP at my request). I have not been able to
> suggest a possible cause based on this.

How are the event_timestamps generated?  If these come from the
transaction timestamp, it is even more weird that tuples seemingly
inserted by the same xmin have differing timestamps.

FWIW I think I've seen similar cases before.

--
Alvaro Herrera                                http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.