Turn a good file into a repeatable contract.
Create a baseline from CSV or Excel, define the structure you expect, and catch schema drift, type changes, missing values, uniqueness failures and duplicate keys before the next import.
Start from a known-good dataset
Define what must stay true
| Key | Column | Expected type | Required | Unique | Max missing % | Max mixed-type % |
|---|
Run a new dataset against the contract
The check validates expected columns, inferred types, missing-value thresholds, uniqueness rules, composite-key duplicates and optional row-count limits.
Manual local checks stay useful. Automation is where Pro begins.
Future paid capabilities can add contract version history, scheduled checks, alerts, CI/API execution, shared team contracts and audit trails without removing the local free workflow.
Catch drift before downstream systems do.
A recurring file can keep the same filename while changing columns, types, null rates or key quality. A data contract turns those assumptions into explicit checks.
The free RowMend contract is stored locally or exported as JSON. No dataset content is sent to a RowMend backend.
Read the practical guide to CSV schema drift and data contracts →