"Christopher Kings-Lynne" <chris.kingslynne@gmail.com> writes:
> see. Collecting the statistics thereafter isn't that hard, but there
> needs to be a way to not collect an exponential volume of statistics on
> all column combinations.
> You could collect them on all FK relationships - is that enough?
As somebody pointed out upthread, collecting these stats on FK
relationships is actually not very interesting: you already know
that the referenced side of the relationship is a unique column,
and so the selectivity stats of the referencing side are enough.
regards, tom lane