Joel Perren writes:
> - generic_restriction_selectivity() returns the default value (0.01) for
> data types that Postgres doesn't collect standard MCV and/or histogram
> statistics for. I think this is what happens with Table B which (quite
> correctly) does not have these statistics in pg_stats
After some further investigation I think I might have actually got this the
wrong way round. I now strongly suspect that what I am seeing is related to
the presence of histogram_bounds for Table A (possibly a bug? or me being
silly).
Looking into the code (
https://github.com/postgres/postgres/blo
