Implement and Manage Data Quality Constraints in Unity Catalog DP-750 Practice Questions (Page 6)

Questions 26–30

1. 26application · medium

   An e-commerce company has a `product_views` table in Unity Catalog. The combination of `product_id` and `view_date` must be unique. The data engineering team uses Lakeflow Spark Declarative Pipelines for ingestion. They want to enforce this uniqueness at the table level and also want to track how many duplicate rows are rejected. What should be done?

   • AAdd a PRIMARY KEY constraint on `(product_id, view_date)`.
   • BUse `DISTINCT` in the ingestion query to remove duplicates before writing.
   • CCreate a view that uses `ROW_NUMBER()` to deduplicate and have all consumers query the view.
   • DAdd a UNIQUE constraint on `(product_id, view_date)` and use a Lakeflow expectation to count violations.
   Reveal Solution Discussion Select an answer first
2. 27application · medium

   A manufacturing company stores machine telemetry in a Delta table. The table schema includes `machine_id`, `temperature`, and `pressure`. The team wants to ensure that any write to the table must have exactly these columns, with `machine_id` not null and `temperature` between 0 and 100. What should they configure?

   • ASet the table property `delta.appendOnly = true` to prevent schema changes.
   • BEnable schema enforcement and add a CHECK constraint for `machine_id IS NOT NULL AND temperature BETWEEN 0 AND 100`.
   • CUse a generated column to fill null `machine_id` with a default value.
   • DUse a Lakeflow expectation with `onViolation = 'warn'` to log any schema mismatches.
   Reveal Solution Discussion Select an answer first
4. 28application · medium

   A healthcare analytics platform stores patient vital signs. The `heart_rate` column must be an integer between 30 and 220, and the `patient_id` column must not be null. The team wants to enforce these rules at the table level so that any invalid write is rejected. What should they create?

   • AA CHECK constraint: `heart_rate BETWEEN 30 AND 220 AND patient_id IS NOT NULL`.
   • BA generated column that casts `heart_rate` to an integer and fills null `patient_id` with 'unknown'.
   • CA view that filters out rows where `heart_rate` is out of range or `patient_id` is null.
   • DA table property `delta.dataSkippingNumIndexedCols` set to 2 to improve validation performance.
   Reveal Solution Discussion Select an answer first
5. 29expert · hard

   A data platform team manages a Unity Catalog table that stores user events. The schema is defined with columns `user_id`, `event_type`, and `timestamp`. A new source will start sending an additional `device_type` column. The team wants to ensure the pipeline continues to run without manual intervention when the new column appears, but they also want to maintain a NOT NULL constraint on `user_id`. What is the best configuration?

   • AUse a view that selects only the known columns and ignore the new column.
   • BEnable schema evolution on the target table and keep the existing NOT NULL constraint on `user_id`.
   • CDisable schema evolution and manually add the `device_type` column before the new source starts.
   • DSet the pipeline to fail when schema drift is detected and alert the team.
   Reveal Solution Discussion Select an answer first
6. 30application · medium

   A financial services company ingests transaction data from multiple sources into a Unity Catalog table. The `amount` column is expected to be DECIMAL(10,2). Recently, a new source started sending `amount` as a string with currency symbols (e.g., '$123.45'). The data engineer needs to ensure the table only accepts numeric values and the ingestion pipeline does not fail. What should be done?

   • AAdd a NOT NULL constraint on the `amount` column and rely on the source to send numeric values.
   • BUse a `CAST(amount AS DECIMAL(10,2))` in the ingestion query and ignore rows that fail.
   • CAdd a CHECK constraint that uses `TRY_CAST(amount AS DECIMAL(10,2)) IS NOT NULL`.
   • DChange the column data type to STRING and let the application layer handle the conversion.
   Reveal Solution Discussion Select an answer first