Teradata Data Warehousing Flashcards

What is a data mart in relation to a Teradata enterprise data warehouse?

Answer: A subject-specific subset of the data warehouse tailored for a particular business unit or function

A data mart is a focused subset of the enterprise warehouse that contains data relevant to a specific department (e.g., Sales, Finance), simplifying access and improving query performance.

What is the role of metadata in a Teradata data warehouse?

Answer: Describes the structure, lineage, and meaning of data to support governance and query optimization

Metadata includes technical (table/column definitions), business (data definitions, ownership), and operational (lineage, freshness) information that supports data understanding, governance, and optimization.

What is the purpose of a staging area in a Teradata ETL process?

Answer: A temporary landing zone where raw extracted data is loaded before transformation and loading into the warehouse

The staging area receives raw data extracts from source systems, enabling data cleansing, deduplication, and transformation before the cleansed data is loaded into the target warehouse tables.

What is referential integrity in a Teradata data warehouse context?

Answer: A constraint ensuring that foreign key values in a fact table correspond to valid primary keys in dimension tables

Referential integrity constraints ensure that every foreign key value in a fact table (e.g., customer_id) matches an existing primary key in the related dimension table, maintaining data consistency.

What is a Type 1 Slowly Changing Dimension?

Answer: The old dimension record is overwritten with the new value, preserving no history

SCD Type 1 simply overwrites the existing dimension row with the updated value, which is simple to implement but loses all historical context of the attribute.

What is the purpose of a bridge table in a Teradata dimensional model?

Answer: Resolves many-to-many relationships between a fact table and a dimension table

A bridge table resolves many-to-many relationships (e.g., a sale associated with multiple promotions) by holding one row per relationship, connecting fact and dimension tables cleanly.