grouping rows -> PCA calculations


Have got a column with cluster identification. Using this column for grouping, would like to perform the PCA on separate sets (i.e. different clusters)

How to do it in KNIME (using looping?)

Use the "Group Loop Start" and pick the column that has the cluster identifiers in it.

Thank you

I hope it will work, but I need some more detailed instructions

My attempt with the sequence:

Group Loop Start->PCA compute->Data to Report

yielded PCA for only the first cluster or, otherwise

"ERROR     Loop End   Execute failed: Input table's structure differs from reference (first iteration) table: different column counts 12 vs. 19"

dont really understand how to activate other loops?

May be, some relevant examples on the Group Loop start node exist for dummies?


Have you checked the "allow variable column types" in the loop end node?




Thank you

It works