IBM Analytics Ideas

This is something that used to work as expected in prior releases, but stopped working since v18.0.0.1 that we are aware of.

Steps taken towards issue:

Step 1: add a Source node
Step 2: add a Sort node
Step 3: select a few fields in Sort node to sort with
Step 4: connect a Filter node between the Source and Sort nodes
Step 5: in the Filter node, rename a field used by the Sort node (ie. rename field3 to field3a)
Step 6: open the Sort node

Observation:
Obs1: In v14.2.0.3 - the field appears red, signifying an invalid field name
Obs2: In v18.0.0.1 - the field is not flagged(highlighted) and looks valid even though it is not.

Obs3: The aggregate node does not have this issue when inserted in this scenario

Step 7: Click Preview

Observation:
The stream will run and complete in either version, and the log messages warn of the missing field.

Example of log messages after renaming "field3" to "field3a" in the filter node:

Information: Stream execution started
Ignore the invalid field "field3" in the Sort node
Stream execution complete, Elapsed=0.09 sec, CPU=0.05 sec

Moving forward:

This seems like something the application should flag as SPSS Modeler does with any other anomaly​.  

We would like to see this corrected in future releases if possible.