Is there a way to correct an inaccurate n_distinct value for extended statistics in PG11?

Hi All,

We have created extended statistics between two columns on a large table. When looking at pg_statistic_ext the value for n_distinct for the combined columns is off of the actual number of distinct values  by approixmately 40%. This is leading to a poor choice of leading tables in some of our joins.

With single column statistics inaccurate n_distinct values can be corrected with "alter table alter column set n_distinct ="

Is there a way to fix incorrect n_distinct values for extended statisitics?

Thanks,

Craig Jackson