validation

Data validation for semantic models involves executing generated metrics, attributes, entities, and domains to perform sanity checks on returned data. It identifies compilation errors, unexpected NULLs, and incorrect magnitudes to ensure data integrity.

Cross-validation compares related semantic model objects to ensure consistency, such as checking filtered metrics against total metrics. This verifies that a subset metric returns values that are logically consistent with the overall total.

You check data integrity by running type-specific validation checks for domains, entities, metrics, and attributes. The process executes these objects to detect logical inconsistencies and verify that returned data falls within expected ranges.

Unexpected NULLs in semantic model metrics are identified during execution and sanity checks. The validation process detects these NULLs alongside compilation errors and incorrect magnitudes, alerting you to potential data inconsistencies or logical issues.

Yes, you can validate newly created metrics by automatically running them to check if values like revenue are positive and within a reasonable range. It also compares the metric against related totals to ensure it is a valid subset.