Semantic Model Data Quality Check

The Semantic Model Data Quality Check task executes from one to twenty rules ordered on a published model and produces a JSON report.

Configuration

For each rule, define:

Field	Description
Name	Understandable label for the check.
Severity	Warning signals the issue; Error may block the workflow.
Query	Semantic query that returns the value to be checked.
Result Field	Column of the first row to compare.
Operator	Equal, not equal, greater/less than, range, null or not null.
Expected Value	Comparison value.
Final Value	Second extreme, required for the range operator.

Additionally configure:

• Fail on error: interrupts the task when at least one rule with Error severity is not met;
• File name: name of the JSON report.

Conceptual example:

Nome: Ordini presenti
Gravità: Error
Campo risultato: totale_ordini
Operatore: maggiore di
Valore atteso: 0

The rule query must return the value to verify in the first row.

Results

The report contains:

• status passed, warning or failed;
• total number of checks;
• number of checks passed, failed and warnings;
• detail of each comparison;
• reference to the generated JSON file.

Use Warning for anomalies to observe and Error for conditions that make data unsuitable for the process. Periodically verify rules when the model or sources change.