MPN vs internal part number: why a BOM needs both

An MPN is what an assembler or distributor can actually buy. It is a poor key for your own records, because a change of source changes it and every BOM that carries it must be edited. An IPN is useless to the assembler, because it means nothing outside your company. Holding both lets your records stay stable while sourcing moves.

The relationship is one-to-many: one IPN can map to several MPNs, which is what an approved manufacturer list (AML) records. The reverse should not happen. Two IPNs for one MPN make duplicate lines and split stock. A distributor's catalogue number is a third kind of identifier; keep it out of both columns.

It needs few columns: the IPN, manufacturer, MPN, a role (preferred or alternate), an approval status and a date or note. Terms differ between companies: some say AML for approved manufacturers and AVL for approved vendors, others use them loosely. Decide whether your BOM holds the base MPN or the full orderable number with its packaging suffix, and apply that to every row. Alternates are covered in alternate and substitute parts.

Some companies build meaning into the number, such as a prefix for the part family and a code for the value. Others use a plain sequence and keep the meaning in the description. Numbers that encode the value are quick to read, but they force a new number, and an edit to every BOM, whenever the value changes. Plain sequences avoid that and need a good description to be useful. Either works if a number is never reused for a different part.

• IPN as the only identifier. The assembler cannot buy from a number only you understand.
• MPN as the internal key. A source change forces an edit to every BOM.
• One header, two meanings. A column called Part Number may hold either, so tools must guess. Label columns Internal PN and MPN.
• Silent substitution. The same IPN carries different MPNs in different BOMs, with no mapping row to explain it.
• Suffix drift. The base MPN in one file and the full orderable number in another, so duplicate checks cannot match them.
• Distributor numbers in the MPN column. They look like part numbers and fail the lookup.
• Look-alike characters. See OCR pitfalls in part numbers.

BOMParser maps headers such as Internal PN, Company PN, Customer PN and SKU to part_number, and Manufacturer Part Number, MPN and Mfr PN to mpn. When a table has a manufacturer column and one plain Part Number column, it reads that column as the MPN and says so in the parser notes. It reports populated lines with no MPN, leaves them out of the upload CSV, and never corrects a part number. When the document has no MPN column at all, the part-number column fills the upload CSV's mpn column and the parser notes say which column was used, so the codes can be checked before the file goes to a distributor. It reads one PDF at a time and keeps no mapping table; hold that in your ERP or a spreadsheet. Convert a file on the BOM to CSV page; pricing lists the plans. Free: one BOM every 24 hours, ten sheets or fewer, no registration. A free run stops at sheet 10 of a longer BOM. Redeem a code and the ceiling rises to 150 sheets.