Horton

Re: progress report for ANALYZE

Robert Haas <robertmhaas@gmail.com>

View thread View on postgr.es
From: Robert Haas <robertmhaas@gmail.com>
To: Tatsuro Yamada <tatsuro.yamada.tf@nttcom.co.jp>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>, Alvaro Herrera <alvherre@2ndquadrant.com>
Date: 2019-07-08T18:10:50Z
Lists: pgsql-hackers 
On Mon, Jul 8, 2019 at 5:29 AM Tatsuro Yamada
<tatsuro.yamada.tf@nttcom.co.jp> wrote:
> 17520|13599|postgres|16387|f|16387|scanning table|4425|4425
> 17520|13599|postgres|16387|f|16387|analyzing sample|0|0
> 17520|13599|postgres|16387|f|16387||0|0  <-- Is it Okay??

Why do we zero out the block numbers when we switch phases?  The
CREATE INDEX progress reporting patch does that kind of thing too, and
it seems like poor behavior to me.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Commits

  1. Doc: Fix and tweak documentation for ANALYZE reporting