Troels Arvin <troels@arvin.dk> writes:
> On Sun, 28 Nov 2004 16:52:47 -0500, Tom Lane wrote:
>> You need to be able
>> to scan the index and identify rows matching a query without making lots
>> of probes into the table.
> But is it cheaper, IO-wise to "jump" around in an index than to go back
> and forth between index and tuple blocks?
Perhaps not --- but why would you be "jumping around"? Wouldn't the
needed info appear in consecutive locations in the index?
Even if that's not the case, the index should be much denser than the
table because it's only storing the keys and not the rest of the
columns. So I'd expect less net I/O even if the access pattern is just
as random.
regards, tom lane