Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.tools -> Re: Database design question
scotch wrote:
> Running Oracle 8.0.5.2 on Solaris 2.7.
> I would like to solicit people's opinions regarding a design aspect of
> one of our production databases.
>
> There are fields within various tables which establish "relationships"
> based upon data within different columns in other tables -- not
> established by keys, and updated by triggers contantanating data and
> using sequences.
>
> Example: Table one has, say, an owner column which would have data
> within it, and based upon that data, it would have a "quasi -
> relationship" to a different column in another table.
> So, for instance Table one would have rows with company123 in it's owner
> column, and say a "quasi-relationship" to Table two would be attempted
> to be established by that table having company123 in say an Account
> column. The Primary Key Data for these tables would then need to match
> as well. This type of "relationship" building is done throughout the
> database.
>
> These columns are updated by triggers which concatanate data and use
> sequences, which are cached.which This seems to me like it could cause
> trouble down the road if there are rollbacks, imports etc..
>
> If this description makes any sense, opinions are welcome.
>
> Thanks in advance
There's one concern regarding sequences. Whenever there's a requirement to
have
"running" numbers (that is, no holes in the range of generated sequnce
numbers), don't cache!
In case of a power failure or other mishap, you loose the cached numbers.
All else, it looks like a generic RDBMS solution, relations being forced via
triggers,
not FK's.
-- Met vriendelijke groet/kind regards, Frank van Bortel Technical consultant OracleReceived on Wed Jan 05 2000 - 00:00:00 CST