Re: More stable query plans via more predictable column statistics

Tomas Vondra <tomas.vondra@2ndquadrant.com> writes:
> On Wed, 2016-03-09 at 12:02 -0300, Alvaro Herrera wrote:
>> Tomas Vondra wrote:
>>> FWIW while looking at the code I noticed that we skip wide varlena
>>> values but not cstrings. Seems a bit suspicious.

>> Uh, can you actually have columns of cstring type?  I don't think you
>> can ...

> Yeah, but then why do we handle that in compute_scalar_stats?

If you're looking at what I think you're looking at, we aren't bothering
because we assume that cstrings won't be very wide.  Since they're not
toastable or compressable, they certainly won't exceed BLCKSZ.
        regards, tom lane