oops...sorry I found Java snippet allows one to create new columns. However, I am still not sure about the "Group by" or "Aggregate by Column" operation. I saw someone had asked the same question and there was a reference to Python Scripting Plugins. It will be great if someone throws some more light on it.
The python scripting nodes help when you want to append a new row or a new column to an existing table. To the best of my knowledge it is not possible to created completely new tables (as you attempt to do in your initial posting).
For instance, the "JPython Script 1:1" allows one to sum up all the values in each of the table's columns and append the result as a new row to the input table.
Sorry for not being more precise. As I pointed out in the thread that you're referring to (I guess), the Jython nodes were not written by the KNIME core team.
Please check out this small R example that might help to handle your problem:
Perhaps a node with an interface for creating derived variables would be a good addition to the platform.
Is there anything like that in the works?
Thanks a bunch Jay, Gabriel and Wiswedel. As Jay pointed out an node with an interface creating derived variables would be awesome. Plus, a node for the 'Group By' operation would be wonderful...I will watch out for KNIME 1.3 :-)
Thanks again Gabriel...the "R Example" was helpful.
Any word on a variable transform node or a grouping node?
Some other comments on grouping capabilities: http://www.knime.org/node/165
Nodes to perform stacking and unstacking of the dataset would also be very useful!