Skip to content

Entity Resolution Recipe FAQ

Common questions about the rule type behind the recipe, its prerequisites, what it writes, AgentQ, and resuming.

The Recipe and the Check

Is the Entity Resolution recipe a new rule type?

No. The recipe is a guided flow around the existing Entity Resolution rule type. It creates an ordinary Entity Resolution check from the fields and threshold you choose, dry-runs it, activates it, scans with it, and helps you act on its anomaly. The check behaves exactly like one created by hand.

Does the check stay active after the recipe finishes?

Yes. The check is created as a Draft when you leave Review, becomes Active when Validate finishes, and stays Active. Future scans keep catching new duplicates, and Remediate in Recipe on their anomalies brings you back to the Remediate step. Archive the check from the datastore page if you no longer want it to run.

Can I load a check I created by hand into the recipe?

Yes. Open in Recipe on a check that is not active yet opens the recipe at Review with the check's configuration, and Scan in Recipe on an active check opens it at Scan. See Entry Points and Resuming.

What happened to "Target Fields"?

The step is now called Compare Fields, and the fields on it are weighted evidence with a comparison type each. Blocking fields, which used to be part of the same step, became an optional Blocking Fields view inside the Distinction Field step. Existing checks kept their behavior and were migrated to the new configuration automatically.

Prerequisites

Why can't I see my table on Select Asset?

Only profiled tables are listed. Run a Profile operation on the datastore and reopen the step. If the table is a computed asset you just created, wait for its profile to finish.

The scan stores the source records of the duplicates it finds in the enrichment destination, and the golden set is written there too. Without a linked enrichment destination there is nowhere for either to go, so the step blocks and offers Link Enrichment, which opens the same dialog as the datastore page.

Outputs

Where is the golden set written, and what is it called?

To the enrichment destination linked to the source datastore, as an output named <prefix>_mat_er_<table>, for example _qualytics_mat_er_ORDERS. When the name needs sanitizing the recipe uses <prefix>_materialize_er_<id> instead. A JDBC destination stores the output as a table. A DFS destination lowercases the name and stores it as a Delta folder, such as _qualytics_mat_er_orders.delta. The exact name is shown on the Materialize step. Inside the enrichment datastore the output also carries the display name <TABLE> / Entity Resolution (<source datastore>).

What is the excluded-records output?

An optional companion output, written when you tick Excluded Records on Materialize, holding the duplicates you did not keep. It has the golden set's name plus an _inverse suffix and the display name <TABLE> / Entity Resolution / Excluded (<source datastore>). It gives you an audit trail of every record the run dropped.

Does the recipe change my source table?

No. The recipe only reads the selected asset. Everything it writes goes to the enrichment destination, and the check, scan, and anomaly live in Qualytics.

Can I keep more than one golden record per entity?

Yes. Tick every record you want to keep in the Keep as golden column. All ticked records go to the golden set, and only the unticked ones are excluded. Each entity needs at least one record kept. AgentQ recommends one record per entity; accepting its pick does not clear the others you ticked.

What happens to the anomaly?

The full scan raises a Shape anomaly when duplicates exist (one per partition scanned), and its source records are the duplicates you remediate. When the golden set is written, the recipe acknowledges that anomaly automatically.

AgentQ

Can I use the recipe without AgentQ?

Yes. Every step works manually: you pick the fields, set the threshold, read the dry-run and scan results, and select the golden records yourself. When no AI provider is configured, the AI banners are simply absent.

Why does AgentQ suggest fields but refuse golden-record recommendations?

The field suggestions, the dry-run analysis, and the scan interpretation send statistics and field names, which the Metadata Shared data sharing level allows. Golden-record recommendations send the records themselves, so they need Source Data Shared. Below that level the step reports that its recommendations could not load and you select the records yourself. See AgentQ Access Controls.

Why does the Remediate step ask before using AI?

Because that step is the only one that sends your data values to the model. Nothing is sent until you click Get AI suggestions.

Resuming and Rerunning

Can I rerun the recipe on new data?

Yes. The check stays active, so after a new load open it on the datastore page and click Scan in Recipe to start at the Scan step. If a scheduled scan already ran, click Remediate in Recipe on the anomaly it raised. Each run replaces the contents of the golden set output for that source table, so it always holds the latest resolution.

Does Start Over delete anything?

No. It clears your progress and returns to the first step. The check stays active on the asset, the scans and anomalies remain available, and the outputs already written to the enrichment destination stay as they are.

Can I resume on another computer?

Not from the Recipes page. Progress is kept in your browser session, so another browser or computer shows Get Started instead of Continue. Use Scan in Recipe on the check or Remediate in Recipe on the anomaly to pick up at the right step, since everything the run produced is stored in Qualytics.