Horton

Re: ERROR: found unexpected null value in index

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Peter Geoghegan <pg@bowt.ie>, Manuel Rigger <rigger.manuel@gmail.com>, PostgreSQL mailing lists <pgsql-bugs@lists.postgresql.org>
Date: 2019-07-16T18:25:54Z
Lists: pgsql-bugs 
Hi,

On 2019-07-11 17:20:07 -0400, Tom Lane wrote:
> I'm a little dissatisfied with the fact that I had to duplicate the
> all-visible checking logic out of nodeIndexonlyscan.c.  Maybe we should
> think about refactoring to avoid multiple copies of that?  But that's
> probably a task for a separate patch, if it's practical at all.

It's maybe worthwhile to note that having the visibilitymap logic in
nodeIndexonlyscan.c. is one of the larger remaining limiting issues from
the tableam work (called out in [1]). We're going to have to move the VM
accesses into the AM at some point not that far away - but it's not
entirely trivial to do so without creating additional VM buffer lookups
/ pins.

Greetings,

Andres Freund

Commits

  1. Fix get_actual_variable_range() to cope with broken HOT chains.

  2. Improve performance of get_actual_variable_range with recently-dead tuples.