View Policy Execution and Results

Every policy in ADOC — Data Quality, Reconciliation, Data Freshness, Data Anomaly, Data Drift, and Schema Drift — runs and produces results, but not all policy types execute the same way. This page explains how execution works for policies that support manual and scheduled runs, including how to trigger an ad-hoc run, how execution modes affect your data, and what the results page shows. For rules, metrics, and result fields specific to a policy type, see that policy's page.

Which Policies Support Ad-Hoc Execution

Data Quality, Reconciliation, and Data Freshness policies can be triggered manually at any time, independent of their configured schedule.

Data Anomaly, Data Drift, and Schema Drift policies execute automatically — Data Anomaly and Data Drift run after each profiling job, and Schema Drift runs after each crawler run. These three policy types cannot be triggered manually.

Why This Matters

Previously, scheduled policies could only run on their configured cadence. Teams needing to validate a data correction, an out-of-cycle load, or a customer escalation had no way to force an immediate check. Ad-Hoc Execution removes that constraint: you can trigger a supported policy manually, at any time, independent of its schedule.

Execution Modes

When you run a supported policy — scheduled or ad-hoc — you choose one of three modes:

1. Full Run – Processes the entire dataset. Does not update the incremental marker.

2. Incremental Run – Processes only new data since the last run. Updates the incremental marker, which future scheduled runs use as their starting point.

3. Selective Run – Processes only user-selected partitions or subsets. Does not update the incremental marker.

Note Running a Full run against a policy normally configured for incremental execution can cause the next scheduled incremental run to reprocess some of the same data, since the incremental marker isn't advanced by the Full run.

Triggering an Ad-Hoc Run

1. Navigate to Manage Policies.

2. Find the policy and click the Play icon.

3. Choose an execution mode: All Your Data (Full), Incremental, or Selective.

4. Optionally select Skip Scoring (see below).

5. Click Execute.

Skip Scoring

Ad-hoc runs can be excluded from scoring so you can validate data without affecting historical trends:

• Skip Scoring selected: The run does not affect asset quality scores and is not factored into future anomaly or threshold calculations. Use this for diagnostic or exploratory runs.

• Skip Scoring not selected: The run's results affect scores and anomaly baselines, the same as a scheduled run. This cannot be undone after the run completes.

Known Limitation Skip Scoring currently excludes a run from direct score changes, but exclusion from relative and anomaly-based calculations is still being validated. A skipped-scoring run may still influence future relative or anomaly comparisons. Until this is fully resolved, treat Skip Scoring as reducing — not guaranteeing zero — impact on relative and anomaly-based results.

Overlap Handling

To prevent duplicate or conflicting executions:

Skipped and blocked runs don't create duplicate compute usage or conflicting incremental markers.

Reading Policy Execution Results

After a run completes, open the policy's Execution Summary page to review:

• Execution History – Start/end time, engine used, execution mode, and rules/metrics configured vs. passed.

• Overall Status – Successful, Errored, Warning, Running, or Skipped, with the reason shown on hover for Skipped runs.

• Execution Details – Rows scanned, processing engine, and filters applied.

• Score / Result Summary – Pass/fail outcome and score impact (reflects Skip Scoring if selected).

The specific tabs and fields shown vary by policy type — for example, Data Quality shows a Segmented Analysis tab, while Data Freshness shows a trend graph of the monitored metric. See the relevant policy page for those details.

What's Next

• Manage Policies – Trigger runs and review the policy table.

• Monitoring and Managing Jobs – See skipped and failed runs across all job types.

• Data Quality Policy, Data Freshness Policy, and Reconciliation Policy pages – Result fields specific to each policy type.