monkragAll notes ↗
← Back to the graph

Put things in order

Unique identifiers

Use stable references to distinguish records and connect them without guessing from names.

2 min readStarter note

A unique identifier is a value that distinguishes one record from others in its defined scope. Customer numbers, invoice numbers and product codes can serve this purpose when their rules are clear and they are maintained consistently.

Names are useful labels, but fragile identifiers. People share names, companies change trading names and product descriptions evolve. A stable reference lets the label change without losing the relationship to earlier records. Some identifiers are unique only within a branch, supplier or year.

A small-business example

Two suppliers both use product code “104”. A purchasing spreadsheet combines their catalogues and matches items on that code alone. Prices and descriptions can now attach to the wrong product.

The business records a supplier identifier alongside the supplier's product code. Together, these identify the supplier's item. It also maintains its own internal product identifier where several suppliers provide the same stock item, with checked mappings between the records.

Try this

Choose a table you frequently combine with another. Name the field, or combination of fields, intended to identify a record. Look for blanks and repeated values. Check whether identifiers survive exports as text, including leading zeroes. Investigate collisions before accepting an automatic match.

Use One row one thing to define the record, Duplicates to examine repeated identifiers, and Provenance to preserve the basis of any manually confirmed mapping.