🔬Granularity

NEW

Also known as: Data Grain, Level of Detail

The level of detail at which data is captured, stored, modeled, or analyzed.

DESCRIPTION

Granularity determines what a single record represents and how metrics roll up. Finer grain improves flexibility but increases storage and complexity. Coarser grain reduces cost and query time but may obscure insights. In dimensional modeling, declaring the grain of facts is foundational. Partitioning and indexing strategies often align to grain. Choosing the right grain balances performance, cost, and analytical needs.

ALTERNATIVE DEFINITIONS

Kimball Group

The grain defines exactly what a fact table’s row represents.

Crisp definition in the context of dimensional modeling.

Source

Wikipedia

Granularity is the extent to which a system is broken down into small parts.

Provides a general systems perspective applicable to data design.

Source

RELATED TERMS

Kimball Methodology Star Schema Feature Scaling Fact Sampling Rate Resolution Indexing Partitioning Summarization

COMPETING TERMS

Aggregation

RELATIONS

• Narrower: Data Granularity

KNOWLEDGE AREAS

Data Modeling Data Warehousing & BI Data Architecture Data Engineering

TARGET AUDIENCE

Data Modeler Data Architect Analytics Engineer BI Specialist Data Engineer

EXTERNAL SOURCES

• Wikipedia - Granularity
• Kimball Group - Authoritative guidance on declaring grain
• Google Cloud - Partition granularity guidance

TOP QUESTIONS

What is Granularity in data modeling?

It’s the level of detail a record represents, defining how data is captured and aggregated.

Category: Definition

Why does Granularity matter?

It affects storage, performance, flexibility, and the insights you can extract.

Category: Relevancy

How do you choose the right Granularity?

Start from business questions, then balance flexibility with performance and cost.

Category: HowTo

What’s a common mistake with Granularity?

Modeling at multiple grains in one fact table causes confusion and errors.

Category: Mistakes

How does Granularity relate to aggregation?

Aggregation summarizes fine-grained data to coarser levels for performance or reporting.

Category: Relation

Can Granularity change over time?

Yes, but changes require careful migration, backfill, and versioning strategies.

Category: Process

What tools help with Granularity decisions?

Dimensional modeling techniques, workload analysis, and partitioning strategies guide selection.

Category: Tooling

How does Granularity impact cost?

Finer grain increases storage and compute but may avoid costly reprocessing later.

Category: Cost

What is the Kimball perspective on Granularity?

Declare the grain first to anchor all dimensional design choices.

Category: Opinion