The Census Bureau provides block relationship files for comparing 2010 and 2020 tabulation blocks. It also publishes tract relationship files connecting the two census vintages. These are separate products with their own layouts and explanations.

Choose the entity type before importing a relationship file. Keep the earlier and later geographic identifiers in separate columns whose names include the vintage. If both are renamed to a generic code, it becomes easy to reverse a join without noticing.

For a crosswalk audit, count distinct identifiers on each side and count rows per identifier. Inspect unmatched records and repeated identifiers before selecting a rule for aggregation or allocation. Record every dropped or combined row as a local transformation.

A geographic relationship is not, by itself, a population allocation instruction. Before applying any weights, identify the documented quantities in the file and the assumptions of the proposed method. Preserve the original relationship rows so another reader can reconstruct the transformation.

This page links to authoritative relationship products. It does not distribute a derived crosswalk, estimate weights or claim that every block or tract has a one-to-one successor.

Sources
[S1] 2020 Geographic Relationship Files — https://www.census.gov/geographies/reference-files/2020/geo/relationship-files.html