Data Warehousing and Business Intelligence MCQs

The load window is constrained by source availability and reporting deadlines.
Performance must be sufficient to finish processing before data is needed.

Drill-through connects an aggregate cell to its underlying detailed rows.
It supports investigation, validation, and operational follow-up.

COUNT DISTINCT removes repeated customer keys before counting.
COUNT(*) would count order rows rather than unique customers.

Hierarchies organize attributes from summarized to detailed levels.
Examples include Year–Quarter–Month–Date and Country–Region–City.

A filtered index stores only rows matching a condition such as Active = 1.
It can be smaller and cheaper to maintain than a full-table index.

A seek navigates the index structure to qualifying keys.
A scan reads a broad portion of an index or table and may be appropriate for large result sets.

Facts store surrogate dimension keys rather than source identifiers.
The lookup resolves the correct dimension member, including its historical version when required.

ELT uses the target system’s scalable processing engine for transformations.
It is common in modern cloud and massively parallel analytical platforms.

Aggregations trade storage and processing time for faster retrieval.
A good design targets common queries and hierarchy paths rather than materializing everything.

SUM adds additive measure values across the selected group.
It is appropriate when each fact row contributes an amount to the total.

A level is one stage of detail within a hierarchy.
Users can roll up from City to State or drill down from Country toward State.

Columnstore organizes and compresses data by column and processes it in batches.
This design is highly effective for warehouse-scale aggregation queries.