Understanding data discrepancy between NDC and annotated OpenFDA data

Question

When comparing the NDC drug data at

https://www.fda.gov/drugs/informationondrugs/ucm142438.htm and OpenFDA annotated data at https://open.fda.gov/apis/drug/ndc/download/ , I observed that the FDA dataset has 150,000+ entries, while NDC has around 130,000 entries.

I also had a look at the relevant code of openfda over here: https://github.com/FDA/openfda/blob/master/openfda/annotation_table/combine_harmonization.py

OpenFDA seems to rely mainly on NDC product.txt , but clearly, it seems to have has more data than NDC.

Could anyone shed more light on why there seems to be data discrepancy?

Answer 1

I'm Jack with the openFDA team. After looking into this briefly, I don't see the issue you are. This may have changed in the intervening time from when you made the post, but I'm currently seeing 154,800 records available through the NDC Directory openFDA API (https://api.fda.gov/drug/ndc.json). The combined total of the product.xls and unfinished_product.xls files is 154,802 (downloaded from https://www.fda.gov/drugs/drug-approvals-and-databases/national-drug-code-directory).

I'll keep an eye on these over the next several updates to ensure that it isn't an inconsistent discrepancy.