The dc3MergeBatch process lets you trigger the Auto Merge of an existing Deduplicate search job from Apex. This is the programmatic equivalent of opening a completed (or previously stopped) duplicate job and choosing Start Auto Merge in its options. You point it at the parent search job, set a score threshold, and it merges every duplicate group in that job that meets the threshold.

This is useful when a search job already ran — perhaps as part of a scheduled batch — and you want to merge its results later, conditionally, from your own automation rather than by hand.

📘

A completed search job is required first

Auto Merge runs against the duplicates found by a search job, so a parent dupcheck__dcJob__c of type search must already exist. See Start a Plauti deduplicate job from Apex to create one from Apex.

🚧

The Batch Merge feature must be licensed

Constructing dupcheck.dc3MergeBatch throws a LicenseException ("Feature not licensed or license expired.") when neither the Batch Merge nor Quick Merge feature is available on the org's Plauti Deduplicate license.

Key Points

• Runs against an existing search job, identified by its dupcheck__dcJob__c record Id — no list of SObjects or record Ids is needed.
• Creates a child dupcheck__dcJob__c record of type merge, linked to the parent search job through dupcheck__parent__c.
• Merges only the duplicate groups whose match score is greater than or equal to dupcheck__autoProcessThreshold__c.
• Executes the merge through the dupcheck.dc3MergeBatch batch class, so it scales to large jobs within Salesforce governor limits.
• Updates the merge job record with the Apex job Id, and marks the parent search job's auto-process type as MERGE, so the operation is traceable from the Plauti Deduplicate UI.

Apex Example

// The Id of the parent search job whose duplicates you want to merge.
// This is a completed (and possibly previously stopped) dupcheck__dcJob__c of type 'search'.
Id searchJobId = 'a0X...';
Integer threshold = 80; // Minimum match score percentage to merge

Savepoint savepoint = Database.setSavepoint();
try {
    // Create a new DC Job record for the merge operation
    dupcheck__dcJob__c mergeJobData = new dupcheck__dcJob__c(
        dupcheck__name__c = 'Merge if score is greater than or equal to ' + String.valueOf(threshold) + '%',
        dupcheck__type__c = 'merge',           // Constant value (do not change)
        dupcheck__sourceObject__c = '001',      // Must match the source object of the search job
        dupcheck__matchObject__c = '001',       // Must match the match object of the search job
        dupcheck__parent__c = searchJobId,      // Link to the parent search job
        dupcheck__autoProcessThreshold__c = threshold,
        dupcheck__autoProcessType__c = 'MERGE'  // Constant value (do not change)
    );

    // Insert the merge job definition
    insert mergeJobData;

    // Start the merge batch process with a batch size of 1
    Id apexJobId = Database.executeBatch(
        new dupcheck.dc3MergeBatch(searchJobId, threshold, mergeJobData.Id),
        1
    );

    // Update the merge DC Job record with the Apex job ID for tracking purposes
    update new dupcheck__dcJob__c(Id = mergeJobData.Id, dupcheck__apex__c = apexJobId);

    // Mark the parent search job's auto-process type as MERGE
    update new dupcheck__dcJob__c(Id = searchJobId, dupcheck__autoProcessType__c = 'MERGE');
} catch (Exception e) {
    // Roll back the job records so a failure does not leave an orphaned merge job behind
    Database.rollback(savepoint);
    throw e;
}

The dc3MergeBatch constructor takes three arguments, in this order:

1. searchJobId — the Id of the parent search job to merge.
2. threshold — the minimum match score percentage.
3. mergeJobData.Id — the Id of the child merge job you just created.

Recommendations

• Use a batch size of 1. A single merge can touch many related records across several objects, so a small batch size keeps each transaction comfortably within governor limits. Increase it only after testing with your own data volumes.
• Wrap the setup in a savepoint and try/catch, as shown above. The DML that creates and links the job records should succeed or roll back together, so a partial failure does not leave an orphaned merge job.
• Set the threshold deliberately. Only duplicate groups scoring at or above the threshold are merged; the rest stay in the job untouched.

When to merge specific records instead

If you do not want a batch-style run over a whole job — for example, you want to merge a known set of records under custom business logic — use the direct merge methods instead. See Merge Methods, which merge a list of SObjects or record Ids you provide.