Posted by Ed Murphy on 04/03/07 05:14

bubbles wrote:

> The tables are in fact seperated by products, as each set of product-
> specific data are handled by
> different engineers. These sets of data need to be sliced and diced
> into various analyses, and the
> engineers would also need to dive into raw data for some of their
> work.
>
> Because these product-specific data sets are large (several million
> records each), I thought that
> perprocessing and seperating them into their respective tables would
> help the engineers access
> their analyses and data faster.

This would probably still be better handled using a single table,
including the separating column(s) in indexes as appropriate - though I
understand the limited motivation for applying further development
effort to a system that does work.

You could save the engineers the trouble of remembering an extra WHERE
clause by providing each of them with a set of views, along the lines of

create view TheTable_eng1 as
select *
from TheTable
where some_column = 'eng1'

[Back to original message]