Create a matcher

A matcher automatically detects duplicate records and groups them into clusters.

Only one matcher can be created for each entity.

Create a matcher

To create a matcher:

  1. Expand the entity node, right-click the Matcher node, and select Define SemQL Matcher….
    The Create New SemQL Matcher wizard opens.

  2. (Optional) In the Description field, enter a description for the matcher.

  3. Click Finish to close the wizard.
    The SemQL Matcher editor opens.

Define the match rules

To define the match rules:

  1. In the matcher, scroll to the Match Rules table.

  2. Click the Add Match Rule button in the Match Rules tables.
    The Match Rule: NewRule editor opens.

  3. In the Name and Description fields (intended for model designers), enter an internal name and description for the match rule.

  4. In the Label and Documentation fields, provide a user-friendly name for the rule and type an informative description of the rule’s purpose, logic, and usage.
    In the documentation, use plain text or the Markdown syntax for rich text. This text provides detailed documentation for the rule. It appears in the documentation side navigation.

  5. In the Color field, specify a color for the match rule. to represent the edges connecting records that match according to this rule. If you do not choose a color, the rule will be assigned a random color.

  6. If you want to use matching on child records, select the Match on Child Records option, and select the Child Records you want to use for matching.

  7. Define the Binning Expressions:

    1. In the Binning Expressions table, click the Add Binning Expression button.
      The SemQL editor opens.

    2. Create a SemQL expression used to bin records for this entity, and then click OK to close the SemQL editor.

    3. Repeat the previous steps to create all your binning expressions.

  8. Define the Match Condition and Match Score properties:

    1. In the Matching section, click the edit expression button Edit Expression button.
      The SemQL editor opens.

    2. Create a SemQL condition to match records for this entity, and then click OK to close the SemQL editor.

    3. Enter a value for the Match Score between 1 and 100.

  9. Press Control+S (or Command+S on macOS) to save the changes in the editor.

  10. Use the breadcrumb navigation at the top of the editor to return to the matcher.
    The new match rule appears in the list.

  11. Repeat these steps to create additional match rules.

  12. Press Control+S (or Command+S on macOS) to save the editor again.

Configure grouping and scoring

To configure grouping and scoring:

  1. In the matcher, scroll to the Grouping and Scoring section.

  2. Choose whether to Use Transitive Match Score. For more information, see direct and transitive scoring.

  3. Choose whether to use Multi-Iteration Grouping. For more information, see multi-iterating grouping.

  4. Press Control+S (or Command+S on macOS) to save the editor.

Configure the merge and auto-confirm policies

To configure the merge and auto-confirm policies:

  1. In the matcher, define the Merge Policy: set the minimum confidence score required for a match group to be merged in the various merge cases.

  2. Define the Auto-Confirm Policy:

    1. Auto-confirm golden records: specify the minimum confidence score required for a match group to be automatically confirmed.

    2. Auto-confirm singletons: select this option to have singletons (i.e., unmatched records) automatically confirmed.

  3. Press Control+S (or Command+S on macOS) to save the editor.