After some wonderful help from Max in getting debugging working with Ant and the hibernate tools tasks I'm now observing behaviour that sees any type of custom reverse engineering strategy I attempt being ignored.

I'm using an Oracle 10g database, and the fields I'm trying to alter are ones that are bizarrely mapped by default to BigDecimals and other weird scenarios.

First I tried the <type-mapping> option in hibernate.reveng.xml - according to the documentation, something like

should result in any mapped numeric field being forced to an integer object. However, it seems to be completely ignored and even any specialisation of the above, adding scale or precision information is equally ignored. I've checked the log4j output on DEBUG and the only messages relating to the fields in question look like:

In oracle, the column in question is configured as a "NUMBER" with no precision or scale.

Secondly I've attempted to implement a custom reverse engineering strategy by implementing DelegatingReverseEngineeringStrategy. My custom strategy is defintely being called (I can see log4j output from it) and is overriding columns by returning "java.lang.Integer", for example, but the resulting mapping files don't reflect the override.

Playing around a little more - adding:

to the hibernate.reveng.xml file above now successfully maps to java.lang.Integer.

Does this suggest the only way to fix the problem is to refine the way all the fields are configured in Oracle?

This still doesn't explain why the Custom reverse engineering strategy implementation is being completely ignored.

if i understand you correctly then what this means is that the sql type returned my your driver for those fields are DECIMAL and not NUMERIC.

remember the jdbc-type is what is defined in java.sql.Types and this does not always correspond to the same names as the db specific name.

(at some point i might add support for db-type or something to allow using db specific names)