Re: proposal : cross-column stats

Dne 13.12.2010 16:34, Tom Lane napsal(a):
> Tomas Vondra <tv(at)fuzzy(dot)cz> writes:
>> Well, until this point we've discussed failure cases involving 'AND'
>> conditions. What about 'OR' conditions? I think the current optimizer
>> computes the selectivity as 's1+s2 - s1*s2' (at least that's what I
>> found in backend/optimizer/path/clausesel.c:630).
>
> If you can solve the AND case, the OR case falls out of that. Just
> replace s1*s2 with a more accurate AND calculation.

Oh yeah, now I see - it's just the usual equation

P(A or B) = P(A) + P(B) - P(A and B)

and we're estimating "P(A and B)" as P(A)*P(B).

regards
Tomas