Followup: Sequence Numbers as Primary Keys -- More Information
Date: 1995/09/26
Message-ID: <44aeb0$spv_at_hudson.lm.com>#1/1
Thanks to everyone who has responded so far. The IT group has compiled more information. But first, a little more background information is necessary.
The project is designed to track/display/plan events in a fast moving environment. Most of the tables do not change very often. A few are updated constantly from TPF/IMS sources as the events take place. Approximately 180 users have to be notified when an update takes place. The IT group has already devised a clever little mechanism to automatically send a message to the users when interesting data comes in. (which BTW, has yet to be approved by the DBAs).
The following points were discussed:
- Speed/Efficiency
2. Insulation from changes in the primary key
IT group claims that using sequence numbers as PK positions them for change. IF PK ever changes, the data distributor would not have to change at all, and the C++ class that contained the changed object would have to be updated. Using composite keys, changes would have to be made to the coding/parsing routines on data distribution as well as changes in the handle objects and the class itself -- not even mentioning the DB changes.
DB group claims that any change would result in modifications to the database and the program anyway.
3. Coding/Development time would be shorter. Less complex code, easier to maintain -- All non-quantifyable right now.
Therefore the question now is ... Is this still worth pursuing? The cost would be to hire IT people who specialize in Oracle DB. The system is critical and has to be up 24x7, so it wouldn't be wise to have only one person support it. There may be some other projects which would require databases, but I don't know if there is enough work for full-time IT DBAs ...
Thanks again for any help,
Joel Hagans
jhagans_at_telerama.lm.com
Received on Tue Sep 26 1995 - 00:00:00 CET