Give spreadsheet rows stable identities before migration

A spreadsheet's row number feels like an identifier until somebody sorts the sheet. If another tab refers to row 73, the relationship may no longer mean what the team expects. Stable record identity is an early migration decision.

Find the real entities

A Colombian operations sheet may mix customers, requests and invoices on one row. Identify which things can exist independently and which relationships must survive. One customer can have several requests; one request may have several documents.

Review duplicates deliberately

Do not merge records only because names match. Keep an original reference and document uncertain matches. A shared company inbox, an abbreviation or a corrected spelling can make automatic assumptions unsafe.

Rehearse a small import

Choose representative rows with a duplicate, missing value and closed request. Check counts and relationships after import, then ask staff to find familiar fictional examples in the new application. Preserve the original file privately until the agreed migration checks pass.

PostgreSQL's constraint documentation explains concepts such as unique keys and relationships. The data model must still reflect your process. A spreadsheet-to-application engagement should include data review and import acceptance rather than only reproducing the spreadsheet as a screen.