Data Management & Governance

Roles

Owner (accountable, decides access) · steward (quality, definitions) · custodian (technical implementation) · user.

Life cycle

Create → store → use → share → archive (retention policy) → destroy (securely, documented).

Relational databases

• Primary key = unique row identifier; foreign key links tables; referential integrity.
• Normalization reduces redundancy. Database management system (DBMS) enforces rules; data dictionary = metadata.
• Structured query language (SQL): SELECT · FROM · WHERE (rows) · JOIN · GROUP BY · HAVING (groups) · ORDER BY.

Storage

Online transaction processing (OLTP) database · data warehouse (online analytical processing (OLAP), cleaned history) · data mart (subset) · data lake (raw, all types).

Controls

Extract, transform, and load (ETL) reconciliations · master data change approval · restricted database access · data lineage.

Exam traps

• WHERE filters rows before grouping; HAVING filters groups after aggregation.
• Normalization reduces redundancy but can slow reporting queries, which is why data warehouses are often denormalized.
• Data under a legal hold must not be destroyed, even when the retention period has passed.