"Steinar H. Gunderson" <[EMAIL PROTECTED]> writes:
> For the record: Could we do with a less-than-optimal split here?
Yeah, I was wondering the same. The code is basically choosing two
"seed" values to drive the index-page split. Intuitively it seems that
"pretty far apart" would be nearly as good as "absolute furthest apart"
for this purpose.
The cost of a less-than-optimal split would be paid on all subsequent
index accesses, though, so it's not clear how far we can afford to go in
this direction.
It's also worth considering that the entire approach is a heuristic,
really --- getting the furthest-apart pair of seeds doesn't guarantee
an optimal split as far as I can see. Maybe there's some totally
different way to do it.
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 3: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faq
