sugi wrote:

>
> I'm wondering if there's an associated performance penalty for using
> BIGINT as opposed to INTEGER for primary/surrogate keys. Would anyone
> have a ballpark figure / estimation of how much slower it would be?
>

None. None at all. Well, if the numbers are random across the range,
you'll get less compression with bigger keys than with smaller, but
that's it. Between prefix and suffix compression, the size of the
underlying field is really not significant.

Regards,


Ann