Match Process

The records with their standardized/derived fields and key values (whether from batch or real-time) will then be passed into the match process, Match – Product Data. This takes one input stream and uses “compare against self” to compare records within the input stream.


The 20 key value array attributes are mapped for clustering. By default each cluster is set with a maximum of 64000 comparisons permitted. If all data is passed in on the working data input, this is approximately equivalent to a maximum cluster size of 250.

Note that these limits apply to Batch Matching. In real-time matching, the external application or hub selects the candidate records for matching based on the key values that are passed back from the key generation service. Any limits in candidate selection are applied externally.