It's still an interesting question.
Why doesn't truncate reset table statistic to 0? The table state prior to a
truncate shouldn't influence vacuum or necessity after the truncate.
Not that vacuuming a recently truncated table would be expensive, but
Analyze timing is a concern.
regards,
Rod
On Mon, Dec 16, 2013 at 10:50 AM, Stephen Frost <sfrost@snowman.net> wrote:
> Anit,
>
> * Anit Chakkarwar (anitchakkarwar@gmail.com) wrote:
> > Now n_tup_del = 2, n_tup_ins=120, n_live_tup=20 in pg_stat_user_tables,
> but
> > how can I figure out what has happened to 98 rows?
>
> This is information for statistics- if you need an accurate count,
> you'll need to use a trigger and track that information explicitly..
> There can be other ways that n_tup_del can end up being inexact.
>
> Thanks,
>
> Stephen
>