View Full Version : kettle: dimension lookup/update step

12-14-2006, 10:43 AM
i'd like to see more flexibility on how the dimension lookup/update step handles the "unknown" row for each dimension. Currently, it creates a surrogate_key = 0 row in the dimension. I'd like to see the ability to pick the surrogate key value and it's meaning. For example, in my dimensions i usually have a few "default" members: key=-1 means "unknown"; key -2 means "not applicable"; key -3 = "hasn't happened yet", etc, etc.

So on creation of dimension, it would insert these rows for me. On lookup of dimension surrogate key in my surrogate key pipeline of fact table processing, i'd have the option of selecting between these "default" members.


05-18-2008, 10:32 PM
Doesn't seem to have been much response on this.

I'd like to see more flexibility here also. Database id's of zero cause problems with certain Java database programming paradigms (in particular EJB3), so we have been trying to avoid them. It seems we cannot avoid them if we want to use Kettle's Dimension lookup/update step.