Answer: Records that violate the expectation cause the job to fail.

## Explanation In Delta Live Tables (DLT), the `ON VIOLATION` clause determines how data quality expectations are enforced. The specific expectation in the question is: ```sql CONSTRAINT valid_timestamp EXPECT (timestamp > '2020-01-01') ON VIOLATION FAIL UPDATE ``` The key part is `ON VIOLATION FAIL UPDATE`. This means: 1. **FAIL**: When any record violates the constraint, the entire pipeline update (job) will fail. 2. **UPDATE**: This specifies that the failure applies to the update operation. Therefore, when a batch contains data that violates this constraint, the entire job fails rather than processing the records in any other way. Let's examine why the other options are incorrect: - **Option A**: Incorrect - This describes behavior similar to `ON VIOLATION DROP ROW`, not `FAIL UPDATE`. - **Option C**: Incorrect - This describes behavior that would require explicit quarantine table configuration, not `FAIL UPDATE`. - **Option D**: Incorrect - This describes behavior similar to `ON VIOLATION RETAIN ROW`, not `FAIL UPDATE`. - **Option E**: Incorrect - This describes behavior that would require additional configuration for flagging invalid records, not `FAIL UPDATE`. In DLT, there are three main violation actions: 1. **FAIL UPDATE**: Pipeline fails when violations occur 2. **DROP ROW**: Violating rows are dropped 3. **RETAIN ROW**: Violating rows are retained but tracked as invalid The highlighted answer (B) in the original text indicates this is the correct answer, which aligns with the `FAIL UPDATE` behavior in Delta Live Tables.

Author: Keng Suppaseth