With this goal in mind, you should ETL directly from your source data to the OMOP CDM. Otherwise, you will have two transformations with the potential for loss during each transformation. One from source to PCORNet and then another from PCORNet to the CDM. If you must go from PCORNet to OMOP, custom mapping is the answer. You will retain the original, PCORNet source value while mapping to a standard concept_id in the OMOP vocabularies. This will allow you to do network research while meeting your institution’s need to retain the source value. I see you have like my other post on custom mapping, but I’ll also link it here for others who read this post.
You might want to ask around the PCORNet community to see if others have a PCORNet to OMOP script.
EDIT: Also, Don is spot on in his above post. Ask yourself if you have a use case for the data. Don’t waste time & resources bringing in data for the “just in case” scenario. You can always bring in more data later. Your source data and the OMOP CDM are continuously evolving.