NULL Vocabulary_versions

Vocabulary Users
1

Good Afternoon,

I wanted to see if anyone could comment on there being NULL Vocabulary_versions. As we are building our v5 product I wrote in requirements that the Vocabulary_Version should not be null. (As the CDMv5 spec’s state). Should this be okay? If so, should we have the spec’s updated? If not, I guess corrections would be made from the OHDSI group?

Issue found on the following:

Thanks,
JG

2

@Jerrod.George:

You have a point. Problem is that these are either home-made or dead (and therefore have no more versions, for example Oxmis). But let’s do the following: We put in the date we touched it last, and that’s the version. If it never changes - so be it.

3

Okay. So just to confirm. Where we find a NULL place in the date we “load” from Athena, and leave it as such. Correct?

4

Yes. But if there is a non-nullable field and Athena gives you a NULL keep screaming. :smile:

5

@Christian_Reich

Thank you sir!

6

In the latest download there are still NULLs in VOCABULARY.VOCABULARY_VERSION column (at least WHERE vocabulary_id=‘Gender’; we are changing DDL to allow nulls - once we load the table, I will report which other vocabularies have NULL in this column).

7

So what’s the decision:
do we change DDL making vocabulary version nullable
or
make put some dummy into the vocabulary, kind of “OMOP generated” or “Unknown”?

8

Ok. so we’re making them nullable in cdm v6.0.
here’s the github issue https://github.com/OHDSI/CommonDataModel/issues/149#event-1871382114

tagging @clairblacketer

1 Like
9

Yes