What causes DELTA_CONSTRAINT_VIOLATION?

Source data contains null values in columns declared as non-nullable in DLT expectations. Unexpected data from an upstream source breaks a domain constraint (e.g., negative prices). A new data source was added without updating the expectation thresholds. The expectation action was set to 'fail' but the threshold is too strict for the data

How do I fix DELTA_CONSTRAINT_VIOLATION?

Step 1: Review the DLT event log to see which expectation failed and what percentage of rows violated it.. Step 2: Inspect the source data for the violating rows — determine if this is a data issue or an overly strict expectation.. Step 3: If the data issue is upstream, fix the source pipeline or transformation before re-running DLT.. Step 4: If the expectation is too strict, update the constraint definition and consider switching to 'warn' or 'drop' action while investigating.. Step 5: Configure a quarantine table to capture and analyze violating rows without stopping the pipeline.

High severitydata_quality

Power BI Refresh Error:
DELTA_CONSTRAINT_VIOLATION

What does this error mean?

A Delta Live Tables expectation (data quality constraint) was violated by incoming rows, and the pipeline is configured to fail rather than drop or warn on invalid records.

Common causes

1Source data contains null values in columns declared as non-nullable in DLT expectations
2Unexpected data from an upstream source breaks a domain constraint (e.g., negative prices)
3A new data source was added without updating the expectation thresholds
4The expectation action was set to 'fail' but the threshold is too strict for the data

How to fix it

1Step 1: Review the DLT event log to see which expectation failed and what percentage of rows violated it.
2Step 2: Inspect the source data for the violating rows — determine if this is a data issue or an overly strict expectation.
3Step 3: If the data issue is upstream, fix the source pipeline or transformation before re-running DLT.
4Step 4: If the expectation is too strict, update the constraint definition and consider switching to 'warn' or 'drop' action while investigating.
5Step 5: Configure a quarantine table to capture and analyze violating rows without stopping the pipeline.

Frequently asked questions

What are the three DLT expectation action modes?

warn logs the violation but keeps all rows; drop removes violating rows and continues; fail stops the pipeline update with an error. Choose based on how critical the data quality rule is.

Can I see which rows violated an expectation?

Yes, if you configure a quarantine table using EXCEPT ROWS WHEN in your DLT flow definition. The DLT event log also shows the violation count and percentage.