Enrico Sirola wrote:
> The application code usually performs selection queries like
>
> select tag, value from test where code='XXX';
>
> also, deletions are like
>
> delete from test where code='XXX';
>
> and insertions follow the same pattern (all the data for a code is
> inserted using a loop in a single transaction). That's more or less all.
>
> so this type of workload is greatly enhanced by an index-organized table
> (oracle) or a clustered index (SQL Server/Sybase).
Hmm - I'm not sure it does benefit that much. I mean, if you're going to
be accessing XXXA, then XXXB, XXXC etc. in order then it clearly helps
to have the table with the same order as your primary key. Otherwise,
I'd be doubtful you'd see that much benefit.
--
Richard Huxton
Archonet Ltd