Entry Points and Resuming
The Recipes page is not the only way into the Entity Resolution recipe. Existing work can pull you in at the right step: a check you are still building opens at Review, an active check opens at Scan, and an anomaly opens at Remediate with its records loaded. This page covers those three entry points and how the recipe keeps your progress between visits.
Entry Points
All three buttons live on the datastore page, on the details of a check or an anomaly. Each appears only for the Entity Resolution rule type and only when you hold the permissions the recipe needs there. Entering the recipe this way replaces any progress you had in it, since the recipe now works on the check or anomaly you opened.
Open in Recipe
Open in Recipe is in the options menu of an Entity Resolution check that is not active yet, such as a Draft. It also appears on an active check when you cannot run operations on the datastore, in place of Scan in Recipe. It loads the check's configuration into the recipe and opens the Review step, with the asset, distinction field, blocking fields, compare fields, and threshold already filled in and every earlier step reachable. The Check badge flashes in Working Assets to show where you came in. From there you continue as usual: Next moves on to Validate, and Validate saves any change you made to the check before running the dry run. See Open a Check in the Recipe.
Scan in Recipe
Scan in Recipe is a button on the details of an Active Entity Resolution check. It loads the check and opens the Scan step directly, skipping the build phase because the check is already validated. The scan does not start on its own; you click Scan when ready. This is the way to rerun the recipe on new data after a fresh load, since the check is still active and its earlier golden set stays in place. See Scan a Check in the Recipe.
Remediate in Recipe
Remediate in Recipe is a button on the details of an anomaly raised by an Entity Resolution check, as long as the anomaly is not archived. It loads the check behind the anomaly and the anomaly's source records, and opens the Remediate step with the duplicates grouped by entity, ready for you to pick golden records. The Anomaly badge flashes in Working Assets. This is how a scan run outside the recipe, whether by hand, on a schedule, or from a Flow, hands its duplicates back to the recipe for remediation. See Remediate an Anomaly in the Recipe.
| Entry point | Where | Shown for | Opens at | Preloaded |
|---|---|---|---|---|
| Open in Recipe | Check options menu | Entity Resolution checks not yet Active, or Active checks when you cannot run operations | Review | The check's configuration |
| Scan in Recipe | Check details | Active Entity Resolution checks | Scan | The check's configuration |
| Remediate in Recipe | Anomaly details | Anomalies of an Entity Resolution check, not archived | Remediate | The check's configuration and the anomaly's source records |
Resuming
The recipe keeps your progress as you work: the step you are on, the configuration, the check it created, and the results of the dry run, the scan, and your golden-record selections. When you leave and come back, the Recipes page shows the Entity Resolution card with an In progress pill and a Continue button, and clicking it reopens the recipe where you left it.
Progress is kept in your browser session, per recipe, rather than as a record in Qualytics. That has two consequences:
- A different browser, another computer, or a cleared session starts the recipe fresh. To pick up the same work there, enter through the check or the anomaly instead: Scan in Recipe reopens an active check at the Scan step, and Remediate in Recipe reopens an anomaly at the Remediate step.
- Everything the run produced persists regardless: the check, the scan, the anomaly, and the outputs in the enrichment destination are ordinary platform objects and stay where they are.
The duplicates loaded for remediation can be large, so they are cached separately in the browser and reloaded when you continue. If they are missing, the Remediate step points you back to Scan to load them again.
Starting Over
Start over in the recipe header, and Start Over on the Complete step, clear your progress and return to the first step. The dialog spells out what that means:
This clears your progress and returns to the first step. Nothing the run already produced is deleted.
- Recipe Progress: your answers are cleared and the recipe reopens at the first step.
- Quality Check: any check the recipe created stays active on the asset.
- Scans and Anomalies: scans already run and the anomalies they reported remain available.
- Materialized Tables: tables already written to the enrichment datastore stay as they are.
Starting over is therefore safe at any point. To remove something the run created, archive or delete it from its own page like any other check, anomaly, or container in the enrichment datastore. See Resume or Start Over.
See Also
-
Examples
Real scenarios showing the recipe taking a table from suspected duplicates to a golden set.
-
Best Practices
Guidelines for preparing the asset, choosing fields, tuning the threshold, and remediating.
-
Permissions
The roles and team permissions behind each step of the recipe.
-
How It Works
The phases, the two prerequisites, what the run reads and writes, and AgentQ's role.
-
Build and Validate
Select Asset, Distinction Field, Compare Fields, Match Threshold, Review, and Validate, step by step.
-
Scan and Remediate
The full scan, choosing golden records, writing the golden set, and the completion summary.