analytics-dictionary / term / transformation & metrics

Data grain

The grain of a table is what a single row represents, such as one order line or one session.

Written by Reflective Data experts · updated Oct 5, 2026

Getting the grain right is the most important modelling decision. Joining tables at different grains is the classic cause of duplicated or inflated numbers.

State the grain in the documentation of every model.

Data modeling Dimension Star schema

// how we can help

We build tested data models and semantic layers, or host them for you. Data build and semantic layers → · Hosted dbt on the platform

Data mart

A data mart is a focused subset of a warehouse built for one team or subject,…

dbt-test.md

dbt test

A dbt test is an automated check that a model meets expectations, such as unique, not…

slowly-changing-dimension.md

Slowly changing dimension (SCD)

A slowly changing dimension is a descriptive attribute, like a customer segment, whose value changes over…

metric-tree.md

Metric tree

A metric tree breaks a top-level metric down into the input metrics that drive it.

dbt.md

dbt

dbt is a tool for transforming data in a warehouse using SQL, with version control, testing…

semantic-layer.md

Semantic layer

A semantic layer defines your business metrics and dimensions once, so every tool and person gets…

← full analytics dictionary