today I made a few changes to a workflow I have been using for a while. To make a long story short, one column needed to be changed from integer to double.
First I did that with the Column Rename node. The output seemed to be okay, but the Z-Primes node from the HCS Tools threw a error message (Execute failed: Could not cast attribute Value (null) to double. Use the String2Number node to convert the type).
Then I tried the Math Formula node to convert the column from integer to double. The node output equals the Column Rename node's output in every little detail. But, guess what! The Z-Primes node works perfectly fine.
I hope this information might help you to identify a possible bug within KNIME.
Have a nice day!
This looks more like a problem in the Z-Primes nodes. The Column Rename just changes the type of the column but does not touch the values itself. This should not affect the downstream nodes, though. They can just use the double representation of the integers in the column. The Math Formula creates a new column with real doubles in it. But it really shouldn't matter for other nodes.
I could reproduce this error. The problem seems to be the convertion of integer to double using the Rename Node. The table actually displayes it as a "double", but if you go to the Spec tab and see the domain, then it shows a "Non-Native" data. I always try to avoid this data type, because I often have problems with it and it seems that the Z-Primes node also has problems handling it and throws this error.
I just tried two KNIME nodes using this "Non-Native" double column and the Math Node has no problem with this data type. The String Manipulator also works, but the column converted to double using Math Node as string gives e.g. "1234.0", whereas the one converted using the Rename Node gives "1234". So it seems that it is not really treated as a double data type as Thor already mentioned, the values stay integers.
I think you should not use the Rename Node to convert integer to double.
We will see if be can fix the Z-primes node to accept this type of columns with the next release...