On Tue, 2012-06-12 at 13:10 -0400, Robert Haas wrote:
> I don't think it's going to solve the problem in general, but TBH I
> don't think Jeff's proposal is, either. I mean, ignoring
> xmin-committed, xmax-committed, and all-visible bits is going to come
> with a pretty steep performance penalty all of its own.
I think we'd certainly have to discourage users from launching lots of
full table scans during a data load. We could go so far as blocking
reads, as I said in my other response, and still (mostly) meet my
primary goals.
But allowing reads without hint bits might be useful for highly
selective index access, which is a substantial use case for the kind of
large tables we'd be bulk-loading.
Regards,Jeff Davis