Date: Tue, 19 Mar 2019 14:00:15 -0700 (MST)
From: legrand legrand <legrand_legrand@hotmail.com>
To: pgsql-hackers@postgresql.org
Message-ID: <1553029215728-0.post@n3.nabble.com>
Subject: [survey] New "Stable" QueryId based on normalized query text
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Precedence: bulk

Hello,

There are many projects that use alternate QueryId 
distinct from the famous pg_stat_statements jumbling algorithm.

https://github.com/postgrespro/aqo
query_hash

https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/AuroraPostgreSQL.Optimize.ViewPlans.html
sql_hash

https://github.com/ossc-db/pg_hint_plan
queryid

Even pg_stat_statement has a normalize function, 
that would answer the current question ...

Here are some *needs* :

needs.1: stable accross different databases,
needs.2: doesn't change after database or object rebuild,
needs.3: search_path / schema independant,
needs.4: pg version independant (as long as possible),
...

and some *normalization rules*:

norm.1: case insensitive
norm.2: blank reduction 
norm.3: hash algoritm ?
norm.4: CURRENT_DATE, CURRENT_TIME, LOCALTIME, LOCALTIMESTAMP not normalized
norm.5: NULL, IS NULL not normalized ?
norm.6: booleans t, f, true, false not normalized
norm.7: order by 1,2 or group by 1,2 should not be normalized
norm.8: pl/pgsql anonymous blocks not normalized
norm.9: comments aware

Do not hesitate to add your thougths
Regards
PAscal


--
Sent from: http://www.postgresql-archive.org/PostgreSQL-hackers-f1928748.html