Trying to use the Joiner node with two Column List Loop Start nodes

KNIME Analytics Platform
1

Hi there, I have to sets of columns (pulled from the same data source using two column filter nodes). The first column filter is pulling pre-test columns, and the second column filter is pulling post-test columns. To keep the example simple lets just say there are 5 pre-test columns and 5 post-test columns that relate to each other.  I want to use the first Column List Loop Start node to pull the first pre-test items and the second Column List Loop Start to pull the first post-test item. I then want to join these two items so that I have a new set of two columns - the first pre-test item and the related first post-test item. I will then run some analysis on that data that will result in a new column (Using the Loop End (Column Append) node) that provides a score based on the pre-post analysis. This would then run through the remaining four items. This would allow me to automate analysis by iterating through each pre-post item set individually and ending up with the resulting scores as new columns.

The problem I run into is that the joiner on the two Column List Loop Start nodes throws an error:

Unable to merge flow object stacks: Conflicting FlowObjects: <Loop Context (Head 0:118:117:146, Tail unassigned)> - iteration 0 vs. <Loop Context (Head 0:118:117:147, Tail unassigned)> - iteration 0 (loops/scopes not properly nested?)

I read an article that uses the Flow Variable to connect the two loop starts but the problem is that it only loops through the columns of one loop start and not the other.

Any ideas?

Thanks,

Stephen

2

Hi Stephen,

You just need to connect the flow variable output port of one Column List Loop Start node with the flow variable input port of another Column List Loop Start node. 

Please find attached sample workflow. 

 

Best,

Anna

3

Anna, thanks for the reply and the example. I did try that, and what happens (as does in your example workflow) is that only one of the nodes cycles through it's columns.  In the example you gave, the first iteration gives you mpg and weight (as expected). However, the second iteration gives you mpg and acceleration, whereas I would want cylinders and acceleration. So close.

Stephen

4

Hey Stephen,

Attached you will find a workflow that does what I think you want to achieve. Not my most elegant creation, but it does the job.

Hope it helps, feel free to ask if there is something unclear.

Cheers,
Ferry