Data Management & Integration Flashcards

A data lake has become a 'data swamp.' Which measure most directly addresses the root cause?

Answer: Implementing metadata management, cataloging, and ingestion governance

Swamps form when data lands without metadata, ownership, or quality controls, and cataloging and governance fix exactly that.

Which architecture combines data lake storage flexibility with data warehouse features such as ACID transactions and schema enforcement?

Answer: Data lakehouse

Lakehouses use open table formats like Delta Lake or Apache Iceberg to add warehouse features on top of lake storage.

What is the main advantage of a canonical data model in an integration layer?

Answer: It provides a common format so each system maps only to the canonical model

A canonical model cuts mapping effort because each system translates to and from one shared format.

Under the CAP theorem, a distributed data store facing a network partition must trade off between which two properties?

Answer: Consistency and availability

During a partition, a distributed system can guarantee either consistency or availability, but not both.

A CDO is asked to prioritize which master data domain to tackle first. Which criterion is most appropriate?

Answer: The domain with the highest business impact and pain from inconsistency

MDM programs succeed when they start where inconsistent data causes measurable business harm.

Which data quality dimension is violated when an order record references a customer ID that does not exist in the customer table?

Answer: Integrity (referential consistency)

An orphaned foreign key breaks referential integrity between related datasets.

What distinguishes an operational data store (ODS) from an enterprise data warehouse?

Answer: An ODS holds current, integrated operational data for near-real-time use, with limited history

An ODS supports operational reporting on current integrated data, while a warehouse keeps long-term history for analytics.