Skip to content

Per-stage results summary for a job run

GET
/jobs/{jobId}/runs/{runId}/results/summary
jobId
required
string

Job definition ID

runId
required
string

Job run ID

workspaceId
required
string

Workspace ID

Run results summary

object
job_id
required

ID of the job definition

string
run_id
required

ID of the job run

string
source
required

Provenance of the summary data

string
Allowed values: stage_artifacts run_aggregates
stages
required

Per-stage summaries; empty when source = run_aggregates

Array<object>
object
stage_index
required

Zero-based index of the stage within the job

number
stage_name

Human-readable stage name, when defined on the job

string
result
required

Overall verdict for the stage

string
Allowed values: MATCHED UNMATCHED
rows_processed
required

Rows processed in this stage

number
rows_matched
required

Rows matched within tolerance in this stage

number
rows_unmatched
required

Rows that did not reconcile in this stage

number
measure_exceed_counts
required

Per-measure tolerance-exceed counts

Array<object>
object
measure
required

Measure name

string
exceed_count
required

Number of rows where this measure exceeded its tolerance

number
column_manifest
required

Column manifest of the stage result artifact, for dynamic UI rendering

Array<object>
object
name
required

Column name as it appears in the result artifact

string
type
required

DuckDB logical type of the column, e.g. ‘VARCHAR’, ‘DOUBLE’, ‘BIGINT’, ‘BOOLEAN’, ‘TIMESTAMP’

string
role
required

Semantic role of the column in the comparison result

string
Allowed values: dimension source_value delta threshold_flag percentage_difference match_flag
measure

Measure name this column derives from (set for measure-derived columns)

string
source_index

Zero-based index of the source connection (set for source_value columns)

number
sample_available
required

Whether a bounded mismatch sample artifact exists for this stage

boolean
sample_row_count

Number of unmatched rows held in the bounded sample

number
sample_size_limit

Configured maximum size of the bounded sample (default 1,000)

number
sample_kind

Which rows the sample parquet contains: ‘unmatched’ (default when absent) or ‘matched_preview’ (stage fully matched; sample is a preview of matched rows)

string
Allowed values: unmatched matched_preview
rows_processed
required

Total rows processed across the run

number
rows_matched
required

Total rows matched across the run

number
rows_unmatched
required

Total rows that did not reconcile across the run

number

WorkspaceId is required

Job run not found