This morning I had launched Catalogue Export script from my R Studio, at the same manner like I do it last one.
But, in main analysis number 430, system give us an error indicating ‘org.postgresql.util.PSQLException: ERROR: relation “cdm5.concept_ancestor” does not exist
Position: 201’.
In our OMOP system, this table always was in VOCAV schema, and this script was executed previously without any error, standing concept_ancestor table in VOCAV schema.
I had must copy table from VOCAB schema to CDM5 schema to run again script.
Somebody had this error too?
Best Regards.
Hi, I’m not sure about the Catalog Export script specifically, but keeping the vocab tables in a separate schema can confuse some tools that expect all the tables to be stored into the CDM schema (which is technically correct, there is no ‘vocabulary schema’ standard, but people refer to the vocabulary tables (ie: that you download from Athena) as the ‘vocabulary schema)’. Others have requested a way to store just a single vocabulary table that services multiple CDMs, and so the notion of a ‘vocabulary schema’ has manifested in some of the tools, but I wouldn’t consider it an ‘official cdm feature’.
So, I make this suggestion to everyone who finds themselves in a vocabulary-separated database: create views in each of your CDMs that point back to the centralized vocabulary schema. This saves you the trouble of duplicating vocabulary tables while maintaining the standard of having the vocabulary tables within the cdm schema.
1 Like
@afernandez
There is indeed bug in the 430.sql script, using the CDM database schema where the vocabulary database schema should be used. Thanks for raising this, I will open an issue there.
Also backing this. e.g. Atlas requires the vocabulary tables to be part of the CDM schema.