Hi,

I have a problem when I try to do reverse engineering on a table that is not physically on the same database I connect to. The database server is running DB2.

The table A from schema S is physically in the db X. There is a link between dbs X and Y, so that table B in schema T on db Y resolves to table A in schema S on db X. In clear text this means:

Database X, schema S, table A
Database Y, schema T, table B => resolves to X.S.A

The problem I have is that I cannot generate anything from the table B, connecting to db Y.

However, there is no problem when I try to do any SQL query on that table with any JDBC-based client: executing the following SQL query on database Y

returns the contents from S.A on database X, as expected.

Is there a known issue, or do I have to set any special options in either hibernate config or the reveng file that I am not aware of? Are aliases and nicknames correctly resolved?

Thank you for your help,
Kariem

i don't know if db2 metadata returns the proper column/table info for those kind of tables.

How can I find out? My first impression was that it doesn't. Any guides on where I could look or ask? Is it possible that this is an issue with the DB driver or DB version? By the way, the version is 8.2, and the driver is current.

Alternatively, I wanted to follow a different approach: generate the meta-data via the database where the table is physically located. At runtime, the application uses the generated mappings/pojos/whatever and does not really need to access the meta-data, which is apparently not provided for these special tables. What do you think?

Of course, it would be interesting to find a solution instead of a workaround.

Thank you,
Kariem

look in the code ;)

we just use normal jdbc metadata calls to get the information. You can tweak that by providing a specialized implementation of a MetaDataDialect.

Look in the docs for MetaDataDialect.