I don't agree.
If we already had a column with the execution plan in the
pg_stat_statements then
it would be a workaround for the problem with the ambiguous names in the
query text column.
But we don't have such column, and I don't want to create a dependency
on the unimplemented feature.
I rather fancy an extra column query_qn with the query text where all
the object names are substituted to the qualified names.
Remember that this problem isn't about only table names. The query may
have views, functions, operands etc. from the 'wrong' schema.
Sergei Agalakov
On 11/27/2018 2:10 PM, legrand legrand wrote:
A part of the answer would be to store
explain (verbose on) select count(*) from t1;
result in pg_stat_statements for the corresponding query...
(Verbose On) gives the "qualified names":
QUERY PLAN
---------------------------------------
Aggregate (cost=19.38..19.39 rows=1 width=8)
Output: count(*)
-> Seq Scan on s1.t1 (cost=0.00..17.50 rows=750 width=0)
Output: a1, a2, a3, a4, a5, a6
Extension pg_store_plans
https://github.com/ossc-db/pg_store_plans
can do it.
Extesion auto_explain can help also.
There are also some tryies to extend pg_stat_statements
with plans see
https://www.postgresql-archive.org/FEATURE-PATCH-pg-stat-statements-with-plans-td5940964.html
Regards
PAscal
--
Sent from: http://www.postgresql-archive.org/PostgreSQL-general-f1843780.html
