Organization changes can rewrite old performance

Regions, managers, customer tiers, and product categories change. A current-only dimension assigns prior events to today’s attributes, making old reports impossible to reproduce. Business teams may need both as-was and as-is views.

Type 1 and Type 2 serve different needs

Type 1 overwrites attributes for corrections or nonhistorical values. Type 2 creates versions with surrogate keys and effective periods. Facts should attach to the correct version by business time; late data and backfills need rematching rules.

Expose time semantics to AI

Describe which date selects a version, whether as-was or as-is is default, and how unmatched or overlapping periods behave. BuildTable can be assessed for relationship and semantic modeling; complex SCD and CDC boundaries require real-data validation.

Regression-test boundary dates

Test the day before, day of, and day after a change, plus late facts and reversals. Keep organization reassignments, product reclassification, and customer-tier changes in every model-release suite.

Public references

Build an AI-ready data foundation

Download BuildTable or talk with us about your data modeling scenario.

Download BuildTable