The work
Tables need to reflect the records the application uses and the relationships between them. We design schemas, identifiers and validation rules, then write the queries or views needed for screens, exports and reports. Work on an existing database can also include query repairs, schema changes and revised access.
A migration starts with the source data. We map fields, identify duplicate or incomplete records and agree how each category will be handled. Repeatable import scripts make it possible to rehearse on a copy and compare the result with the source.
Counts alone cannot confirm a transfer. We also reconcile relevant totals and trace selected records through their relationships. The transfer plan covers the last source changes, records held for review and the conditions for restoring the previous state.
Technical considerations
- A migration needs an authoritative source and a decision on how conflicting values will be handled.
- The transfer plan should define when source data stops changing, how the result is checked and how to recover if checks fail.