Bug or OOO: define variable attributes not working properly, or not implemented, when VLM is present (i.e., define_variable_ccode)

[mhungria]

Bug or OOO for the core engine v1.0: define variable attributes not working or not implemented for VLM (Value Level Metadata).

At least define_variable_ccode is returning null when there is VLM defined. It means that apparently it disregards any VLM defined.

See example with draft rule CG0999 running as a local rule, using the released DatasetJSON v1.1 example on git: https://github.com/cdisc-org/DataExchange-DatasetJson/tree/master/examples/sdtm

Note:
I am testing with the core engine v0.14.1 on a linux machine.

[OGarcia11]

@mhungria is this currently blocking an official Conformance Rule? If so, which one?

[alexfurmenkov]

hi, @mhungria. here some updates on this issue - it was possible to extract define_variable_ccode for some of the variables from define.xml for LBORRESU and LBSTRESU by finding code using ValueListDef->CodeListRef . There is no code for LBSTRESC, so error is still there.
here is new report: CORE-Report-2026-02-11T19-08-48.xlsx
datasets used:
dm.json
lb.json
suppdm.json
test_dataset.json

Rule: rule.yml

[mhungria]

Hi @alexfurmenkov , thanks for the update.
Would you run a test with a variation of the yml check, replacing the "not_equal_to" with "equal_to"?
In that way, we can see the list of ccodes from VLM when there is more than one. We would be interested in getting the list of those ccodes in the output, not just one.

Check:
all:
- name: variable_name
operator: equal_to
value: library_variable_name
- name: variable_name
operator: equal_to
value: define_variable_name
- name: library_variable_ccode
operator: non_empty
- name: library_variable_ccode
operator: not_equal_to
value: define_variable_ccode

[alexfurmenkov]

@mhungria , new report with the equal_to instead of not_equal_to condition.
CORE-Report-2026-02-12T10-27-34.xlsx
can you please clarify, what do you mean by

getting the list of those ccodes in the output

[mhungria]

@alexfurmenkov , thanks for the test changing the CG0099 logic, I just wanted to see what was the output like and since you didn't have any difference with the CG0099 logic, I suggested to run a test adjusting the logic for me to visualize the output.

Could you add the following TS dataset to the test? It should provide negative results with the CG0099 not_equal_to logic. It should list all the different ccodes referenced for TSVAL, VLM.

ts.json

[alexfurmenkov]

@mhungria report with ts.json
CORE-Report-2026-02-15T18-31-42.xlsx
this is how were metadata variables filled