cdisc-org / cdisc-org/cdisc-rules-engine
Data merge by variable with uncertainity
- Dominant language
- Python
- Stars
- 113
- Forks
- 43
- Avg merge
- 14h 51m
- Merged PRs (30d)
- 14
Description
When I look into PMDA rules, I see the following.
SD0065
USUBJID/VISIT/VISITNUM values do not match SV domain data
'All Unique Subject Identifier (USUBJID) + Visit Name (VISIT) + Visit Number (VISITNUM) combination values in data should be present in the Subject Visits (SV) domain. There is an exception for records, where Status (--STAT) variable value is populated. E.g., --STAT = 'NOT DONE'. Another exception is for records, where Occurrence (ECOCCUR) is 'N' in EC domain.
I believe that simple implementation is to merge dataset with SV dataset by USUBJID, VISIT and VISITNUM including left-join condition. But VISIT/VISITNUM variables are not always included in datasets. Is such situation, what rule description is preferrable? I can specify USUBJID/VISIT/VISITUM in "Match Dataset" section. However, error is reported for datasets that don't have VISIT/VISITNUM variables. (error: Column not found in data)
Contributor guide
Assessment
This issue has not been assessed yet.