`aip_sdk.DataQualityRun`

aip_sdk.DataQualityRun(id: str, version_id: str, checked_at: datetime, status: str, row_count: int, checks: list[dict[str, Any]], label: str | None, triggered_by: str, run_status: str = 'UNKNOWN', created_at: datetime | None = None, selected_checks: list[str] = list[str]())

A single historical data-quality check execution for a dataset version.

Returned by Dataset.check_runs(). Mirrors the canonical data_operation_runs API resource: one row per execution rather than overwriting a single blob.

Attributes

id str: 16-char run ID.
version_id str: The dataset version this run belongs to.
checked_at datetime: When the checks were evaluated.
status str: Aggregated data verdict: "PASS" / "WARN" / "FAIL" / "ERROR". "ERROR" means an op did not execute cleanly, so the data verdict is inconclusive.
run_status str: Run lifecycle, distinct from the data verdict: "PENDING" / "RUNNING" / "COMPLETED" / "FAILED". Use this to tell "still running" / "errored" apart from the data status.
row_count int: Number of rows in the dataset at check time.
checks list[dict[str, Any]]: Raw list of NamedCheckResult dicts.
label str | None: Optional user-supplied name (e.g. "pre-mapping"); auto-generated "Run N" when absent.
triggered_by str: Who initiated the run: "sdk" / "api" / "ui".
created_at datetime | None: When the run record was created.
selected_checks list[str]: The checks this run set out to run. Every check the dataset's schema admitted, unless the request named a subset. Empty for runs recorded before the platform captured the set.