From: | Christoph Dalitz <christoph(dot)dalitz(at)hs-niederrhein(dot)de> |
---|---|
To: | pimentel_ligia(at)hotmail(dot)com |
Cc: | MuK(dot)Rudolph(at)t-online(dot)de, pgsql-general(at)postgresql(dot)org |
Subject: | Re: Table with 90 columns |
Date: | 2002-09-16 14:15:44 |
Message-ID: | 20020916161544.569d6f88.christoph.dalitz@hs-niederrhein.de |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
>
> Ligia Pimentel schrieb:
> > Yes, a very wide table (many columns) will be less efficient than a table
> > with less columns (this is a matter of relational concepts and
> > normalization).
> >
I must admit that I do not understand this comment:
what has normalization to do with performance?
As far as I understand normalization it is meant to avoid *redundance* and
not to improve performance. Actually normalization in general decreases
performance becaus a join over several tables is much less efficient than a
select on a single table. In most (but not all!) cases it is more important
to avoid inconsistencies due to redundance however.
The only disadvantage of a table with many columns that I can imagine occurs
when the columns are addressed by name rather than by index. If postgres
implements the column lookup by a linear search through all column names,
this can become an issue in very wide tables.
Even in that case I do not know whether a breakup in n tables might increase
performance.
Christoph Dalitz
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2002-09-16 14:21:53 | Re: mod_auth_pgsql |
Previous Message | Oliver Neumann | 2002-09-16 14:02:11 | STored Procedures |