Developer Transformation Guide

10.4.0
- 10.5.2
- 10.5
- 10.4.1

Back Next

Cluster Output Options

Select a cluster output option when you want to organize similar or identical records in the output data.

When you select a cluster output option, the transformation adds a cluster ID value to each output record. You can sort the records by the cluster ID values. The transformation output includes a row for every record. If a record does not match another record with a score that meets the match threshold, the transformation assigns a unique cluster ID to the record. Use the

Match Output

view to select or update the cluster output options.

You can select the following cluster output options:

Clusters: Select the option to assign cluster ID values to the output records.
Clusters - Best Match: Select the option to add the record pair with the highest match score to a cluster. Because a record might represent the best match with more than one other record, more than one record pair can share a cluster ID value.
Clusters - Match All: The
Clusters - Match All
option works in the same way as the
Clusters
option.; The transformation uses
Clusters - Match All
and
Clusters - Best Match
as option names in identity match analysis.

If a Data Integration Service runs multiple Match transformations concurrently, the Data Integration Service generates unique cluster ID values for the output from each transformation. Therefore, the cluster ID values for the records that each transformation generates can be non-consecutive.

The Clusters Option and the Clusters - Match All Option

Select the Clusters option in field match analysis. Select the Clusters - Match All option in identity match analysis.

The Match transformation uses the following rules to create the clusters:

When two records have a match score that meets the match threshold, the Match transformation adds the records to a cluster.

When a record in the data set matches any record in the cluster, the transformation adds the record to the cluster.

If a record in one cluster matches a record in another cluster, the process merges the clusters.

The transformation performs a continual sweep of the match results until all the records belong to a cluster.

If a record does not match any other record in the data set, the transformation assigns a unique cluster ID value to the record.

The Clusters - Best Match Option

Select the Clusters - Best Match option in identity match analysis.

The transformation uses the following rules to create the clusters:

The transformation identifies the record that has the highest match score with the current record. If the match score meets the threshold, the transformation adds the pair of records to a cluster.

If one of the matching records is in a cluster, the transformation adds the other record to the current cluster.

The transformation performs a continual sweep of the match score results until all the records belong to a cluster.

A cluster can contain a single record if the record does not match any other record in the data.

You can use the

Match

property on the

Match Output

view to specify how the transformation compares a single data source to a persistent data store. The

Match

property determines whether the transformation looks for duplicates within the source data or the persistent data store.

Rename Saved Search

Table of Contents

Developer Transformation Guide

Developer Transformation Guide

Cluster Output Options

Cluster Output Options

The Clusters Option and the Clusters - Match All Option

The Clusters - Best Match Option