> It will be really useful to see some test results where you alter the
> fillfactor and report various measurables.
It's not that easy... stressing "only" the index insertion
speed won't be simple. I would have liked some "theory"...
The docs seem to imply there are some guidelines, it's
just that it's too cryptic:
"for heavily updated tables a smaller fillfactor is better
to minimize the need for page splits"
"heavily updated" -> does it mean tables that are inserted/updated
or only "updated"???
"leaf pages are filled to this percentage [...] when extending the index
at the right (adding new largest key values)."
Does it mean that since I will (almost) always add new largest key
values, I should have a big or small FILLFACTOR???
I know that theory is one thing and real testing another; but I can't
test everything; if there are some (proved?) guidelines I'd like to
use them (example: I'm not going to test that fillfactor in table creation
in my case won't make any difference in performance; I trust the
docs and the fact that "it makes sense").