Hello everyone,

I can’t find a solution to a tricky problem, that I guess needs to be addressed with a recursive loop combined with a chunk loop, but despite several attempts I’ve been unable to configure it. The solution must use just native KNIME nodes, no R, no Python.

Here’s the issue:

Imagine you have a matrix with K columns (where K can change) and N rows (where N can change). The values in the matrix are floats. For example, something like this (4 columns, 14 rows):

For each pair of columns, I have to apply a transformation that changes the initial values, updates the values in the starting matrix, and continues until a pre-defined number of iterations.

Using the matrix in the example, since I have 4 columns, there are 6 possible combinations (AB, AC, AD, BC, BD, CD) and the steps must be as follows:

Loop 1, six steps:

- select columns A and B → values transformation → update the values of A and B
- select columns A and C → values transformation → update the values of A and C
- select columns A and D → values transformation → update the values of A and D
- select columns B and C → values transformation → update the values of B and C
- select columns B and D → values transformation → update the values of B and D
- select columns C and D → values transformation → update the values of C and D

Loop 2: I repeat the steps of loop 1 with the modified values.

Loop 3 - Loop n: same steps repeated multiple times

Any idea?

Thanks

Daniele