Why matching commercial vehicle OEM part numbers in SQL is messy (and how we open-sourced our parser)

Tom7668h — Sat, 24 Oct 2026 21:06:55 +0000

When building an automotive or commercial vehicle price comparison engine, the biggest data bottleneck isn't scraping or API rate limits—it's inconsistent OEM and OES part numbers across different supplier feeds.

While building the catalog pipeline for CrazyPartSnatcher (a real-time price comparison platform for commercial vehicle, truck, and trailer spare parts), we constantly ran into the same matching issue across our database tables.

The Problem: 1 Part, 5 Different String Formats

A single truck brake pad, sensor, or filter from manufacturers like MAN, Mercedes-Benz Trucks, DAF, Bosch, Hella, Febi Bilstein, or DT Spare Parts shows up differently in almost every CSV feed and marketplace listing:

Supplier A: 81.50110-0144
Supplier B: 81 50110 0144
Marketplace listing: 81501100144
Mercedes OEM format: A 000 420 00 20 vs A0004200020
Bosch 10-digit format: 0 986 479 R66

If a workshop mechanic searches for 81.50110.0144 and your SQL query only checks the raw string—or strips characters inconsistently—you miss valid supplier offers.

The Solution: `@crazypartsnatcher/oem-normalizer`

Instead of keeping this regex and formatting logic locked inside our private backend, we extracted it into a lightweight, zero-dependency TypeScript/JavaScript package and published it on npm and GitHub.

It handles four core tasks out of the box:

normalizeOem(input): Strips formatting noise (dots, dashes, spaces, slashes) into a strict uppercase indexable string for fast SQL/MariaDB lookups.
generateSearchVariants(input): Generates common vendor-specific spacing and punctuation variants for fuzzy search queries.
detectManufacturerFormat(input): Identifies probable OEM/Aftermarket numbering schemas (e.g., Bosch 10-digit, Mercedes A-numbers, MAN formats).
buildPartSearchUrl(oemNumber): Builds direct lookup URLs for live price checks.

Quick Usage Example

Install via npm:

npm install @crazypartsnatcher/oem-normalizer

Use it in your Node.js / TypeScript pipeline or frontend search:

import { 
  normalizeOem, 
  generateSearchVariants, 
  detectManufacturerFormat,
  buildPartSearchUrl 
} from '@crazypartsnatcher/oem-normalizer';

const rawInput = '81.50110-0144';
const cleanKey = normalizeOem(rawInput); // "81501100144"
const variants = generateSearchVariants(rawInput);
const formats = detectManufacturerFormat(rawInput);
const comparisonUrl = buildPartSearchUrl(cleanKey);

Live Interactive Playground & Source Code

We also built a browser-based interactive tester on GitHub Pages where you can paste truck and trailer OEM numbers to test the normalizer and jump straight to live aftermarket price comparisons for brands like Bosch, Hella, Brembo, ATE, Mahle, and Febi Bilstein:

Live Production Search: CrazyPartSnatcher – Nutzfahrzeug & Lkw-Ersatzteile Preisvergleich
Interactive OEM Tester (GitHub Pages): OEM Part Normalizer Docs & Playground
npm Package: @crazypartsnatcher/oem-normalizer on npm
GitHub Repository: Tom7668h/oem-part-normalizer

If you are working with TecDoc data, automotive catalogs, or supplier feeds, feel free to fork the repo or open a PR with additional manufacturer regex patterns!

Tom7668h / oem-part-normalizer

Zero-dependency TypeScript/JS library to normalize, parse, and cross-reference commercial vehicle & truck OEM part numbers (MAN, Mercedes-Benz, DAF, Bosch, Hella, Febi). Official core utility powering the CrazyPartSnatcher Nutzfahrzeug & Lkw-Ersatzteile price comparison engine.

@crazypartsnatcher/oem-normalizer

A lightweight, zero-dependency TypeScript utility library for normalizing, formatting, validating, and generating search variants of automotive and commercial vehicle OEM, OES, and aftermarket part numbers.

This open-source engine serves as the core part number normalizer behind CrazyPartSnatcher – Preisvergleich für Nutzfahrzeug- & Lkw-Ersatzteile, helping workshops, fleet operators, and automotive developers bridge formatting discrepancies across catalogs, ERP inventories, and multi-supplier databases.

The Problem: The Automotive Part Number Chaos

In automotive and commercial vehicle parts data, the identical physical replacement component is represented across ERP systems, TecDoc catalogs, supplier CSV feeds, and workshop management software in dozens of incompatible formats:

MAN Truck & Bus: 81.50110.0144 vs. 81501100144 vs. 81.50110-0144 vs. 81 50110 0144
Mercedes-Benz / Daimler Truck: A 000 420 00 20 vs. A0004200020 vs. 0004200020 vs. 000 420 00 20
Bosch: 0 281 002 461 vs. 0281002461 vs. 0-281-002-461
DT Spare Parts: 4.61234 vs. 461234
…

View on GitHub

DEV Community: Tom7668h