Get a data diff summary
Structured counts for a completed data diff: primary-key stats, value stats, and — for
same-database diffs — a schema section.
The response shape depends on the diff’s state and engine, so branch on status and on
which sections are present rather than assuming a fixed payload:
running— onlystatus. (The model also permitspending; the server does not currently emit it.)error—statuspluserror.successon a same-database diff —ApiDataDiffSummaryForDone, includingschema.successon a cross-database diff —ApiCrossDataDiffSummaryForDone, which has noschemasection. Schema drift is not reported on this endpoint for cross-database diffs (nor bydiff_stats.diff_schema, which is 0 for all of them); read it via the MCP toolsget_datadiff_overvieworget_datadiff_result_section(section="columns").
Authorizations
Use the 'Authorization' header with the format 'Key '
Path Parameters
Response
Successful Response
- ApiDataDiffSummaryForDone
- ApiCrossDataDiffSummaryForDone
- ApiDataDiffSummaryForFailed
- ApiDataDiffSummaryForRunning
- InternalApiDataDiffDependencies
Same-database (join, or unset algorithm) summary of a completed diff. Carries the
schema section; its cross-database sibling ApiCrossDataDiffSummaryForDone does not —
see that model for what is and isn't reported there.
Results of the diff, materialized into tables.
Schema-comparison counts. Carried by ApiDataDiffSummaryForDone (same-database
diffs) only — the cross-database variant has no schema field, see
ApiCrossDataDiffSummaryForDone.
Scope: column presence, column order, and data types. Nullability and other column
constraints are not compared, so a column that only dropped NOT NULL counts as
matching here.
done, success When the diff was created.
Data source of dataset A ("t1").
Data source of dataset B ("t2").
Columns excluded from the comparison.
Distribution statistics for exclusive rows (rows that exist in one table but not the other). Includes histograms and enum distributions for primary key columns. Only available for diffs computed after October 2025.
File URI for dataset A, when compared as a file.
File URI for dataset B, when compared as a file.
Row filter applied to dataset A.
Row filter applied to dataset B.
Columns compared (all shared columns if empty).
Primary key column names used for this diff.
SQL for dataset A, when compared as a query.
SQL for dataset B, when compared as a query.
Sampling ratio; null when the diff was not sampled.
Path segments of dataset A, when a table.
Path segments of dataset B, when a table.
