Re: Normalization of utility queries in pg_stat_statements

Hi Michael!

I'm rebasing a patch "Tracking statements entry timestamp in
pg_stat_statements" for applying after this patch. I've noted that
current tests are not quite independent one from another. There is two
statements in the end of user_activity.sql test:

DROP ROLE regress_stats_user1;
DROP ROLE regress_stats_user2;

Those are done after the last pg_stat_statements_reset call in this
test file and thus, those are included in checks of wal.out file:

query
| calls | rows | wal_bytes_generated | wal_records_generated |
wal_records_ge_rows
-----------------------------------------------------------------------
-------+-------+------+---------------------+-----------------------+--
-------------------
DELETE FROM pgss_wal_tab WHERE a > $1
| 1 | 1 | t | t | t
DROP ROLE regress_stats_user1
| 1 | 0 | t | t | t
DROP ROLE regress_stats_user2
| 1 | 0 | t | t | t

Those statements is not related to any WAL tests. It seems a little bit
incorrect to me.

Are we need some changes here?
--
Andrei Zubkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company