Hello everyone. I’m going through the ETL process of source value to OMOP CDM. I would like to use my source key as person_id but it is 64 bit so does not fit within an integer. Will have to make it big int. Will I have any problems running Achilles for QA if I cutomize person_id to bigint? Any other advice on how to handle this? No I do not want to place it in person source value and use random generated person_id. Thank you.
Hello @muniras, welcome to the community.
It’s OK to extend person_id from int to bigint; the same applies to string fields’ length. In fact, you might want to choose bigint for all primary identifier fields in OMOP CDM tables (person_id, condition_occurrence_id, drug_exposure_id, etc.) for consistency and scalability. This will not cause any problems with Achilles / ATLAS.
Make sure not to use bigint for concept_id fields, these must remain 32-bit integers.