At 10:29 +0300 on 28/06/1999, Ansley, Michael wrote:
> >>> you "index the living daylights out of" so that the non db saavy
> managers
> >>> who want to do ungodly joints and sorts on tables for organizational
> >>> reporting get decent performance.
> I'm so pleased to find out that somebody else has picked this up. In fact,
> the summary tables that I'm working on are a kind of mini-warehouse, the
> just happen to reside in the same tablespace as the transactional tables.
> However, data warehousing is quite an important issue. I know that most of
> the people who use PG work on transactional systems, but if anyone tries to
> run even a small warehouse on PG it's going to get complicated very quickly.
Maybe I'm missing the point here, but it seems to me that if you simply use
indices not as key definitions but as query accelerators (as in "index the
living daylights out of"), then you may as well define a separate index on
each and every field. Why do multiple-field indices in such a case?
The case for defining uniqueness is a good case.
Other than that, we should remember that indices are merely a method of
acceleration of queries. There will not me much of a difference in
efficiency of queries if the index is split into the individual fields.
When postgres finally has built-in referential integrity, the question may
rise again.
Herouth
--
Herouth Maoz, Internet developer.
Open University of Israel - Telem project
http://telem.openu.ac.il/~herutma
