Re: CASE Survey
Date: 31 Jan 94 19:57:35 GMT
Message-ID: <Jan.31.14.57.34.1994.7822_at_andromeda.rutgers.edu>
deboer_at_wsl34 (Gert de Boer) writes:
>Richard D Holowczak (holowcza_at_andromeda.rutgers.edu) wrote:
>: dagmar_at_indian.mitre.org (Dagmar A. Bogan) writes:
[stuff deleted]
>Can you explain why this happens?
This is a known behavior of Oracle CASE. When you make the move from analysis to implementation, CASE looks at your analysis information and then it looks at your existing implementation (i.e. tables). From a clean slate, with no tables implemented, all design and analysis information is conisdered and all relationships are implemented as you have seen. In this case, the implementation follows directly from the analysis.
However, if you delete portions of the implementation (such as the table def. you deleted), this is not reflected in the analysis. When you go forward again to implement the new tables, some existing table definitions are found so they are left alone. Often this is done at the expense of leaving out foreign keys.
The reason for this is as follows. Lets say I define a table, and then use this table in modules, etc. Then, someone else comes along and adds a table with a relationship to my table which will cause a FK to be added to my table. CASE will not do this. If you look at the impact analysis reports, they will tell you the consequences of these actions.
So, what is the solution ? When deleting a table, all other related tables must also be deleted all the way down to "source" entities, those with no FKs. Then these can all be recreated properly in one step.
Note that if analysis is done properly, this should only be required in extreme situations.
Rich Holowczak
Rutgers University
holowcza_at_andromeda.rutgers.edu
>_______________________________________________________
>Gert de Boer
>Swiss federal institute for forest, snow and
>landscape research (WSL)
Received on Mon Jan 31 1994 - 20:57:35 CET