On Dec 10, 2007, at 1:21 PM, Erik Jones wrote:
> You beat me to the punch on this one. I was wanting to use modulo
> operations for bin style partitioning as well, but this makes things
> pretty awkward as well as unintuitive. So, to the postgres gurus:
> What are the limitations of check constraints when used with
> constraint exclusion? Is this really the intended behavior?
>
/me too!
I have vague recollection of reading that the constraints on the child
tables needed to be free of computation (ie, just straight comparison
ranges) but I can't find that reference now.
But in my case, I can almost always pick the appropriate sub-table
from the application level anyway.
