Bruno Wolff III wrote:
> On Thu, Jun 02, 2005 at 18:00:17 +0100,
> Richard Huxton <dev@archonet.com> wrote:
>
>>Certainly, but if the text in the logfile row is the same, then hashing
>>isn't going to make a blind bit of difference. That's the root of my
>>concern, and something only Dinesh knows.
>
>
> Sure it is. Because the hash can be used in the primary key instead of
> of the error message which should reduce the size of the key enough
> that he can use a btree index.
Sorry - obviously not being clear. Since he's using the index via a
primary key he'll need the columns that key is over to be unique. If the
columns fail that test in the real world, hashing will replace the
index-size error with an "unable to insert duplicates" error.
--
Richard Huxton
Archonet Ltd