Interactive Pattern

Historical Correction

Historical Correction preserves corrected business history without losing what was previously known.

Problem

Corrected history can conflict with previously published reports.

Source systems frequently correct past data. A contract may receive a backdated change. A customer attribute may be corrected months later. A policy status may be updated retroactively.

Historical reporting must decide whether reports should show the corrected truth or the information that was known at the time.

Changing past reportsAudit disagreementLost reproducibilityInvisible retroactive changes

Interactive Pattern

A January report shows Retail. In March, a correction says January should have been Premium.

Choose how the model handles the correction. Each strategy answers a different reporting question.

Correction timeline

Jan

Feb

Mar

Apr

Report published

Correction arrives

Known in January

Customer SegmentRetail

Corrected in March

January SegmentPremium

Correction strategy

Preserve valid time and visible time

January as-known shows Retail. Corrected truth shows Premium.

The correction is valid for January but visible from March. Both reporting perspectives remain explainable.

Trade-off: More complex, but reproducible.

Reporting question

Should a rebuilt January report show Retail because that was known in January, or Premium because that is the corrected business truth?

Why it happens

Business truth and reporting knowledge do not always change at the same time.

A correction may be valid for a past business period, but only become visible to the reporting system later. Without explicit correction handling, rebuilt reports can look different from the reports originally published.

Backdated changesCorrected master dataPolicy restatementsLate source fixesVisible-time gapsAudit requirements

Common modeling approaches

Preserve both corrected truth and historical knowledge.

Bitemporal modeling

Track both valid time and visible time so corrections can be placed on the correct timelines.

As-known reporting

Rebuild reports using only information that was visible at the reporting time.

Persisted snapshots

Store published report outputs when exact reproducibility is required.

Version retention

Keep previous knowledge states instead of overwriting corrected history in place.

Validation checks

Measure how corrections affect historical reporting.

✓ Detect retroactive source changes✓ Validate report reproducibility✓ Track visible-time history✓ Compare current truth vs historical knowledge✓ Measure correction impact on published reports

Why it matters

Historical corrections are a main reason to move beyond simple SCD2.

Without correction tracking, it becomes impossible to explain why reports generated in the past differ from reports generated today.

Historical Correction introduces the distinction between business truth and reporting knowledge.

Related Patterns

Snapshot Reproducibility Historical Conformance Historical Coverage Gap Historical Overlap Dimension Completion

Try it

Explore correction and reporting-time behavior in the Workbench.

Use the Historical Modeling Workbench to reason about corrected history, visible-time logic, historical joins and reproducible reporting.

Open Historical Modeling Workbench →