Glossary · Data, timing and communication
Data validation
In systems engineering, data validation is the set of checks that confirm data meets defined structural, semantic, range, timing and quality rules before it is used. It tests the data, not whether a system fulfills its intended use.
- Data and interfaces
- Integrated systems
- Verification
In one sentence
Data validation checks structure, meaning, range, timing and quality of data before use, and it must define what happens when data fails.
Example
A robot cell controller rejects a recipe from the MES whose gripper force is outside the permitted range and keeps the last approved recipe active.
How it applies
- Integrated systems: Derive the rules from the schema and the interface specification: data types and required fields (structure), units and meaning (semantics), limits (range), age and order (timing), and the quality status delivered by the source.
- Functional safety: Define the reaction to invalid data, such as rejecting it, using a safe default or stopping. A check without a defined reaction does not reduce risk.
- Technical documentation: Operator and integration manuals should explain validation messages and what the user is expected to do about them.
- AI and retrieval: Data that feeds AI models or documentation pipelines needs the same checks. Unvalidated inputs carry errors into generated content.
Data validation vs. validation
Validation in systems engineering confirms that a system fulfills its intended use. Data validation is a runtime or import check on individual data. The shared word causes confusion in requirements and test reports, so always qualify it.