How do you optimize the request for the number of values in the columns of the table by hundreds of millions?

[kalena]

How can I optimize the request to obtain the number of values in the table columns?

My request now looks like this:

SELECT count('col1'), [...] count('col25')
FROM scheme.table

Is he right, is he in any way optimized or rewritten differently?

The problem is, it's supposed to be in the table with which I work 158 million lines, and I'm afraid it's gonna be done just before I die.

[inna]

Maybe you can. approximate statistics that can be found https://www.postgresql.org/docs/12/view-pg-stats.html :

select * 
from pg_stats
where schemaname='<your_schema_name>'
  and tablename='<your_table_name>'

It may be possible to establish indices for non-small columns and compare the compliance plan with and without indices. But to do so carefully - sometimes the creation of an index can hurt. This is particularly relevant for the tables that are often written, since postgreSQL will have to update the indices.

---

The PS can certainly be remarked. https://swarm64.com/post/postgresql-columnstore-index-intro/ but according to the commentaries, you have no right of admin.