On 06/25/07 09:58, Tom Lane wrote:
[snip]
>
> The fly in the ointment is that if the column value is so high
> cardinality as all that, it's questionable whether you want an index
> search at all rather than just seqscanning; and it's definite that
> the index access cost will be only a fraction of the heap access cost.
> So the prospects for actual net performance gain are a lot less than
> the index-size argument makes them look.
Well they definitely are for data warehouses, in which many
high-cardinality columns each have an index.
Because of their small disk size, ANDing them is fast and winnows
down the result set. That's the theory, of course.
--
Ron Johnson, Jr.
Jefferson LA USA
Give a man a fish, and he eats for a day.
Hit him with a fish, and he goes away for good!
