Creating a Rebuild Match Tables Job

Learn how to create a Rebuild Match Tables job in Tenant Management, including the optional FERN vectorization step.

The Rebuild Match Tables job recalculates potential matches for a tenant based on the current match rules. Run this job when you need potential matches to reflect recent changes; for example, after loading data or updating match rules.

If your tenant uses FERN for AI-powered matching, the Rebuild Match Tables job includes vectorization by default to update vector data before the rebuild starts. You can turn off vectorization if vectors are already up to date.

Prerequisites

Before you create a Rebuild Match Tables job, make sure you have the following permissions and access:

Permission to update the L3 tenant configuration using the Configuration API
ROLE_ADMIN_TENANT role to run the Rebuild Match Table task from Console
Access to the Tenant Management application in Console

To create a Rebuild Match Tables job:

In the Console, select the Tenant Management application.
Click NEW JOB.

The Create new job page displays all the jobs that you can run.
Select the Rebuild match table job to view to view the Rebuild match table page.

Specify values for the job parameters based on your requirement.


Parameter	Description	Default
Entity type	Select the entity type, or select All types to process every entity type.	All types
Skip entities count	Number of entities to skip before the job starts processing.	0
Entities limit	Number of entities to process. Leave blank to process all entities.	All
AI-powered vectorization	Turn on to run FERN vectorization before the rebuild. The AI-powered vectorization parameter appears only if your tenant uses FERN. This parameter is on by default. When on, the job runs vectorization to update your records for AI-powered matching. The rebuild starts only after vectorization completes successfully. When off, the job rebuilds the match table without running vectorization.	On
Delete Old Column Families	Turn on to delete existing column families as part of this run.	Off
Distributed	Turn on to split work across multiple parallel jobs on different API nodes.	Off
Job parts count	Number of parallel jobs for distributed mode. Available only when Distributed is on. Reltio recommends setting this to the number of API nodes available to run the job.	2

Click RUN JOB.

Result

The job rebuilds the match table for the selected entities based on the current match rules, so that new or changed data is considered during matching. The result depends on the parameters you selected:

AI-powered vectorization on: The job checks whether vectorization is needed. If needed, the job runs vectorization first and starts the rebuild only after vectorization completes successfully. If not needed, the job proceeds directly to the rebuild.
AI-powered vectorization off: The job rebuilds the match table without running vectorization.
Delete Old Column Families on: The job deletes existing column families before rebuilding. Distributed on: The job splits work across the number of parts set in Job parts count.

Verification steps

Check the job status in Tenant Management to confirm the job completed successfully.
If AI-powered vectorization was on and vectorization ran, check the Completed tab for these jobs in execution order:
1. AI vectorization task
2. AI vectorization worker task
3. Rebuild match table
Review potential matches in Hub to confirm they reflect the current match rules and, if applicable, current vector data.

For API parameter names and additional options not available in Console, see Rebuild Match Table Task.

Unify and manage your data

Creating a Rebuild Match Tables Job

Footer