Build a Crosswalk Between PMID, PMCID, and DOI

Keep the identifier type with the value

Create separate text columns for PMID, PMCID, DOI, and source row. An unlabelled number can be ambiguous. PMCIDs use the PMC prefix, and the numeric portion can coincide with a PMID for a different record. Preserve the input exactly as received before producing a standardized lookup value.

Submit a single identifier type per request to the PMC ID Converter and retain the returned record or error for each input. If the workflow depends on a particular manuscript or article version, examine the converter's version option and the article landing page. A missing mapping may simply mean that the work falls outside the converter's coverage.

Reconcile a hypothetical import

A bibliography contains 50 rows but only 46 distinct PMIDs. The converter maps 38 distinct PMIDs to PMC articles. Eight distinct inputs still need another retrieval route. Recording 38/50 as the identifier match fraction would mix a distinct-identifier numerator with a source-row denominator. The distinct-PMID fraction here is 38/46, approximately 82.6%.

Join verified mappings back to every source row and check whether the join preserved the original 50 rows. Inspect any citation with conflicting DOI or title metadata before overwriting an identifier. Keep the raw mapping response and retrieval date so a later update can distinguish a newly archived article from a change in the local matching logic.

PMC ID Converter API documentation

Related worksheet: Database join planning and audit worksheet.

Source checked on 5 November 2026.