Mappings, Executors, Copy to result etc..

[Dan]

Hi,

So I'm chucking this out there, hoping that some people on the product side read this mailing list.  Otherwise I'm not really sure where these sorts of discussions should go?

Having just built a heavily metadata driven system we built mappings to handle a standard way of reading various levels of metadata. These then get used by our main process in various places, they're normal mappings, not "simple" ones as they have no input. All good.

However we now built another use case where we need to iterate over our readers several times. Great! Transformation executor ahoy then...  However; to get the results out, you have to use "copy files to result". And I didn't want to put that in the mapping, just for this use case - so I created ANOTHER level, which is simply a transformation which executes the mapping, and copies the rows to result.

That seems a bit naff though.. Could it understand that mapping output when using a trans executor == copy rows to result?

(Looking at this from a etl dev perspective, not a kettle dev of course.)

Thanks!

Dan