Data validation
Checking that data meets defined rules, correct type, allowed values, required fields, before it is trusted, entered, or imported.
Data validation is the gate that keeps bad values out. The rules can be simple (a quantity must be a positive number, an email must contain an @) or structural (every row must have an order ID, dates must fall in a range). In spreadsheets, Excel's Data Validation feature enforces some of this at entry time with dropdowns and constraints.
Validation is most valuable at boundaries, when data is entered, imported, or handed between systems, because errors caught there are cheap and errors caught later are expensive. It pairs naturally with comparison: validating a file's structure first (right columns, right types) means a later diff is comparing clean data, not chasing problems that are really validation failures.