I have a new database to handle. The schema is generated by a (java) persistence layer with no constraints implemented. constraints are checked by the persisitence layer within the application.
In a way you could say in this project database design does (almost) not exist. Please don't blame me - no way to change that
Now I am looking for a way to visualze the schema anyway to get kind of a control. I was thinking about doing the following:
- implement informational constraints. Nobody can blame me then for doublechecking constraints. And the persistence layer can be told to generate them.
- Use a tool to visualize the schema with those informational constraints.
questions:
1) do you guys think that is actually possible ?
2) which tools to use for this re-engineering ? I think my boss won't buy me a rational licence. I have no Erwin either, any suggestions are welcome.
I want to pump in the schema-DDL and get the modell as a nice graph.
3) I have heard db2look can generate the schema using the TeX option, but I did not get it working on windows and could not test. Did anybody actually try it ?
Oh well, any help is appreciated.
It is just tough to go first through the object-modell layer and see which classes are related when checking for missing indexes and doing any kind of performance tuning. Also to check that the persistence modell actually makes sense (checking 1..1 relations etc.) is rather difficult since nothing is automated for me from the modeller side ...
In a way you could say in this project database design does (almost) not exist. Please don't blame me - no way to change that
Now I am looking for a way to visualze the schema anyway to get kind of a control. I was thinking about doing the following:
- implement informational constraints. Nobody can blame me then for doublechecking constraints. And the persistence layer can be told to generate them.
- Use a tool to visualize the schema with those informational constraints.
questions:
1) do you guys think that is actually possible ?
2) which tools to use for this re-engineering ? I think my boss won't buy me a rational licence. I have no Erwin either, any suggestions are welcome.
I want to pump in the schema-DDL and get the modell as a nice graph.
3) I have heard db2look can generate the schema using the TeX option, but I did not get it working on windows and could not test. Did anybody actually try it ?
Oh well, any help is appreciated.
It is just tough to go first through the object-modell layer and see which classes are related when checking for missing indexes and doing any kind of performance tuning. Also to check that the persistence modell actually makes sense (checking 1..1 relations etc.) is rather difficult since nothing is automated for me from the modeller side ...