SI Store Ops

Inspectable example · updated 2026-10-11

Synthetic supplier-key matrix: zeros, long codes, reused SKUs and an unmapped item

Four invented product mappings expose numeric coercion and cross-supplier collisions; an unknown key stays unassigned instead of becoming a guessed product.

An example, not a customer case study. Scope and evidence limitations are described below.

The invented mapping

Alder and Birch are made-up supplier labels. The example's declared policy uses supplier plus exact text SKU; it does not trim, case-fold or turn a code into a number. The internal product IDs are invented and contain no people or client records.

If the matrix is wider than the box, scroll horizontally to read every column. Keyboard: focus the matrix and use Left/Right.

supplier | SKU text          | expected internal product
alder    | 00123             | part-a
alder    | 123               | part-b
birch    | 00123             | part-c
alder    | 9007199254740993  | part-d
birch    | 123               | unmapped; no automatic creation

Expected checks

The first two rows stay separate despite looking numeric. The first and third stay separate despite an equal local SKU. The long code remains the exact string shown. The last lookup is unmapped and reports an exception. These are expected results under the authored policy, not observations from a supplier import.

  • A numeric 123 supplied where a text SKU is required is an invalid input in this example.
  • No existing record is changed to satisfy the mapping.
  • A real supplier's rules for spaces or letter case must be agreed separately.

What it does not prove

The batch's local policy model tests these invented lookups. It is not a CSV parser, spreadsheet application, product-equivalence engine or database constraint test. It does not show that an upstream spreadsheet preserved the original source identifier, or that a supplier mapping is correct for physical goods. Do not infer production reliability or previous SI delivery.

  • Recover already damaged identifiers from an authoritative source rather than padding them by guesswork.
  • Keep unknown product mapping under the purchasing owner's approval.

Use this to specify a priced enquiry

Describe which synthetic lookup fails and the desired key. One existing local matching defect may fit fix-one-bug-with-regression-test, from £295 after bounded reproduction and a fixed repair quote. Historical catalogue cleanup, a new master-data system and schema migration are outside the repair. No code, real price lists, private exports or credentials in first contact; the price is untested and payment follows passing checks and sign-off.

Sources and limits