Marko Kreen wrote:
4-byte xids on btree may create data corruption.
Can you be more specific on this? I'm aware of xid being an unsigned
integer which means we need to deal with the cast resulting in negative
numbers. This means we have to split our range queries into several
pieces when
I've been working with Matt on this. Thanks for the suggestions.
Greg Stark wrote:
On Thu, Jun 11, 2009 at 11:25 AM, Matt Amos wrote:
what problems are we going to create for ourselves if we create a
btree index on xmin casted to int4? would it be as efficient to use a
hash index, create a
