I was working on importing the Linguistic variants and found some things that I would like to confirm are accurate or not.
Consider LOINC 2.54 “LOINC_2.54_fr_BE_23_LinguisticVariant.csv” and these two lines:
986: “11218-5”,“Albumine”,“Concentration de masse”,“Temps ponctuel”,“Urine”,“Quantitatif”,“Seuil de détection <= 20 mg/L bâtonnet diagnostique”,“Chimie”,“”,“”,“Tigettes diagnostiques.Seuil détection <=20mg/L”
and
4619: “15267-8”,“(Acer negundo+Betula verrucosa+Fagus grandifolia+Quercus alba+Juglans californica)
-made change to common term - pls confirm Anticorps.IgE”,“Seuil”…
Line 986 has an entity for the < in RELATEDNAMES2, but not METHOD_TYP. Why? I’d expect both to be an entity or both literal.
Line 4619 contains an embedded newline between “californica)” and “-made”. This happens in more files, sometimes preceeded by a carriage return entity �D; (16 in it_IT and 401 in zh_CN).
When I look at the embedded newline they almost always seem unintentional and caused by a user entry error. As far as the entities go I’d really appreciate it if there were either no entities or all entities. No entities seems to make most sense since entities are really not expected in plain text.
Thank you for bringing these issues to our attention. I can confirm that the issues raised to exist in the 2.54 release of the LOINC linguistic Variants. We have started and internal review with the goal of having these problems addressed in the new release.
I apologize for any inconvenience caused by these errors.
Thanks. I’ve been able to work around all issues (I think) so I’m glad to say there’s no immediate need for a fix on my behalf. Nonetheless the more straightforward the import process is in the future, the better it serves the community at large.
