The challenge
Moving business data from spreadsheets into Dataverse, Microsoft's business data platform, is risky: large files, lookups to resolve, duplicates, and runs that fail halfway. An occupational health system had to be migrated without overwriting what was already there.
What we built
We built a .NET 8 migration engine that maps spreadsheet columns to Dataverse values, resolves lookups to related records, detects duplicates within the file and validates each row against declared rules. Batches run in parallel within a set limit, so memory stays steady regardless of file size.
Interrupted imports resume where they stopped, and add-on modules can supply custom transformations and checks. Operators get a dry-run mode, confirmation prompts showing record count and target, and success, error and summary files for every run. A companion tool defines the Dataverse tables in code.
Outcome
Used for a production migration of an occupational health system in October 2026. The migration included a pre-flight schema check, record-by-record verification with aggregate counts compared, removal of reviewed duplicates, and a rule that a migration never overwrites an existing employee's name.
The tool was packaged to run on a machine without .NET installed.