|From:||Klaudie Willis <Klaudie(dot)Willis(at)protonmail(dot)com>|
|To:||David Rowley <dgrowleyml(at)gmail(dot)com>|
|Cc:||PostgreSQL mailing lists <pgsql-bugs(at)lists(dot)postgresql(dot)org>|
|Subject:||Re: BUG #16521: n_distinct_inherited does not affect child partitions when set on main partition|
|Views:||Raw Message | Whole Thread | Download mbox | Resend email|
Thanks for you thorough answer!
"n_distinct affects the statistics for the table itself, while n_distinct_inherited affects the statistics
gathered for the table plus its inheritance children."
I see that this text has multiple interpretations. I read it more like: while n_distinct_inherited affects the statistics gathered [by analyze] for the table plus [affects the statistics of] its inheritance children.
The core problem for me was that I had to set n_distinct manually on my individual partitions for the planner to choose a sane plan for queries on the declared partition. The data for a column was so skewed that the analyzers n_distinct was off by a factor of 1000X, which made all partitions do a table scan instead of index scan. In my mind, setting the declarative partition variable n_distinct_inherited would solve that; but obviously I am wrong on that.
|Next Message||PG Bug reporting form||2020-07-02 03:16:27||BUG #16523: The wal logs fill the disk and are not cleaned up|
|Previous Message||Tom Lane||2020-07-01 16:47:19||Re: [bug] Table not have typarray when created by single user mode|