Sample questions
Data Quality ValidationDifficulty 1
What does "schema validation" mean in a data pipeline?
- aEncrypting every column before it is written to storage
- bCompressing records to reduce the storage footprint of a table
- cChecking that incoming data matches the expected structure✓
- dRenaming columns automatically to match a naming convention
Explanation:Schema validation checks that incoming records conform to an expected structure — field names present, types correct, required fields non-missing — before the data is trusted downstream. It is a structural check, not encryption, compression, or renaming.
Data Quality ValidationDifficulty 2
At which point in a pipeline is schema validation most useful?
- aAs early as possible, right when data enters the pipeline✓
- bOnly at the very end, right before a dashboard reads the final table
- cOnly inside the analytics team's reporting queries
- dOnly once a year during an annual data audit
Explanation:Validating as close to ingestion as possible prevents malformed data from propagating through every downstream stage, which is cheaper to fix than catching the problem after many transformations have already consumed the bad data.
Data Quality ValidationDifficulty 1
What is a "data contract" between a data producer and a data consumer?
- aA legal document signed once a year covering all company data
- bAn explicit agreement about the shape and guarantees of the data a producer delivers✓
- cA physical storage quota assigned to each team's database
- dA billing agreement about cloud infrastructure costs
Explanation:A data contract is an explicit, often machine-checkable agreement between the team producing data and the team(s) consuming it, covering schema, semantics, and delivery guarantees — so changes on the producer side don't silently break consumers.
Data Quality ValidationDifficulty 2
What main problem does a data contract help prevent?
- aSlow query performance caused by missing indexes
- bHigh cloud storage bills from keeping too much historical data
- cNetwork latency between two data centers
- dA producer silently changing data shape or meaning✓
Explanation:The core value of a data contract is catching producer/consumer breakage early: if the producer changes a field's type, meaning, or removes a field, the contract makes that visible instead of letting it silently corrupt downstream consumers.
Data Quality ValidationDifficulty 1
What does a "null check" on a column look for?
- aWhether values are missing in a required column✓
- bWhether the column name contains the literal word "null"
- cWhether the table has zero rows
- dWhether the column's data type is a string
Explanation:A null check verifies that required fields are actually populated, flagging rows where a value is unexpectedly missing — a basic completeness safeguard.
Data Quality ValidationDifficulty 2
What does a "duplicate check" typically detect?
- aRows where a numeric column contains a negative value
- bColumns that have been renamed since the last pipeline run
- cMultiple rows representing the same logical record✓
- dTables that have not been queried in the last 30 days
Explanation:A duplicate check looks for rows that logically represent the same entity appearing more than once, typically identified by a unique key or business identifier repeating — which can silently inflate counts and sums downstream.