"Qingqing Zhou" <[EMAIL PROTECTED]> writes:
> "Tom Lane" <[EMAIL PROTECTED]> wrote
>> No, that's completely irrelevant to his problem. The reason we can't do
>> this is that the transformation from "x << const" to a range check on x
>> is a plan-time transformation; there's no mechanism in place t
"Tom Lane" <[EMAIL PROTECTED]> wrote
>
> No, that's completely irrelevant to his problem. The reason we can't do
> this is that the transformation from "x << const" to a range check on x
> is a plan-time transformation; there's no mechanism in place to do it
> at runtime. This is not easy to fix
"Qingqing Zhou" <[EMAIL PROTECTED]> writes:
> "Robert Edmonds" <[EMAIL PROTECTED]> wrote
>> Instead of specifying explicit address ranges in the query, I'd like
>> to store the ranges in a table:
> Good illustration. I guess we have a problem of the historgram statistical
> information.
No, that
"Robert Edmonds" <[EMAIL PROTECTED]> wrote
>
> EXPLAIN ANALYZE
> SELECT *
> FROM inet_addresses
> WHERE addr << inet('10.2.0.0/24')
>OR addr << inet('10.4.0.0/24')
>OR addr << inet('10.8.0.0/24');
>
> Bitmap Heap Scan on inet_addresses (cost=6.51..324.48 rows=1792335
> width=11) (actual