So, 2E can generate DDL for your tables instead of DDS? I wish Plex would do that. I'm working on a similar problem here. I want to assign long names to our fields in addition to our existing 10 char impl names. That way, JDBC/ODBC interfaces can use the long name. This is mainly for query tools such as MS Access. I have figured out that assigning a long name to a field does not change the format identifier for a file. So, you can recreate the file with the long names and not have to recompile any of your existing RPG programs.
Dean, Crispin,
this is exactly what we do while modernizing Synon built applications and DDS described databases on iSeries machines. To avoid most of the stupid and boring work that has to be done, if you want to put your Synon datamodel to Plex, we developed (and still do) our own tool, to do this work for us. We exactly follow what you described above, but we have an automated process, that triggers the generation of the tables and after this modifies the table source file, putting all the neccessary column renaming into it.
So we actually can use Synon-views and the existing RPG programs and on the other side use our new JAVA classes with JDBC access. Nothing has to be re-generated or recompiled. And as an additional benefit, the system uses the new SQL engine (since r6.1) which is much faster, because we created the tables DDL-described.
Because we put all the triples defining the data model into Plex using the Plex API-model, using Synon model files or the information in a Database schema, we found it neccessary to always know the original objectnames and the Synon implnames (ddsvnm) and so store it all in the Plex model. This is, to also be able, finding the way back to the original object for changes.
I would like CA to build some new triples for defining more informations on DDL defined database tables, so that the power of the iSeries SQL engine can be used much better.
Thanks for sharing your thoughts.
Axel