Hi all,

we stumbled over a interesting behavior regarding Pan and shared objects.

If a transformation contains objects, e.g. data base connections, that are also defined in the shared objects file, Pan overwrites their definitions with the definitions from the shared objects file during execution.

As I understood it, shared objects are meant to ease development in Spoon by providing a way to reuse objects. When saving a transformation in Spoon, the shared objects definitions are written to the .ktr just as any other object. Therefore the transformation becomes independent from the shared object file and can be executed on any other machine.

When using Pan, I would expect a transformation to be executed as is, meaning as defined in the .ktr file. I would not expect Pan to do any subtitution or overwriting of objects with definitions from the shared objects file.

I would like to put up two questions discussion:

1. Is it a desired bahavior of Pan to use the shared objects file?

2. If yes, does it make sense? Why should Pan need do that, since the .ktr files contain all needed information to execute a transformation with Pan?

Thanks for any input.

All the best,
Ingo