Glossary · System coordination, integration and orchestration
Data transformation
Also known as: data conversion
German: Datentransformation
In data integration, data transformation is the conversion of data from one structure, format, unit or granularity into another, for example by mapping fields, converting units, aggregating values or enriching records with data from other sources.
- System integration
In one sentence
Data transformation converts data between structures, formats, units or granularity by mapping, converting, aggregating or enriching it.
Example
Raw 10 ms spindle-load samples are transformed into one-minute averages with order and tool IDs before they are sent to the cloud analytics service.
How it applies
- Integration: Transformations run in integration layers, edge gateways or pipelines. Keep them deterministic and testable, and log which transformation version produced each output.
- Data quality: Aggregation and filtering remove information. Document what is lost, for example peaks inside an averaging window, so that later users don't draw conclusions the data cannot support.
- Documentation: Transformation rules belong in the interface or data documentation next to the Data mapping. For data used in AI training or regulatory evidence, keep lineage from raw to transformed data.
Data transformation vs. data validation
Data validation checks whether data meets defined rules and rejects or flags violations. Transformation changes the data. Validate before and after transforming, because transformations can introduce errors.