It seems that a batch of ICD10 codes where added to the vocabulary with mixed case. This was later ‘corrected’ by deprecating the concepts with mixed case and replacing them with all upper case ICD10 codes. However, at least in the October 2016 V4.5 source to concept map table the result is that both the mixed case version and upper case version of the same code produce different results
select source_code, source_vocabulary_id, target_concept_id
from vocabulary_4_5_oct_16.source_to_concept_map
where lower( source_code) = ‘t81.4xxd’
AND invalid_reason IS NULL
source_code source_vocabulary_id target_concept_id
T81.4xxD 70 0
T81.4XXD 70 437474
The result for SQL Server, where for many sites string matches are case insensitive, is an inconsistency. Has this been corrected?