Spectrum Technology Platform

 View Only

Matching Within a Single Data Source Using Multiple Match Keys

  • 1.  Matching Within a Single Data Source Using Multiple Match Keys

    Employee
    Posted 01-07-2020 15:51
      |   view attached

    In Single Customer View (SCV) implementations we often match customers within the input data file for the initial load and each daily delta update and assign a unique party Id to records representing the same customer and create a best of breed master customer record. Also, on each daily delta update, we match these master customer records from the input customer feed with the existing master customers' in the master customer repository. During each match stage, we also often use Match Keys to flag and compare only records that are likely to match based on criteria such as similar first and last name or the same postal codes because we don't want to compare unlikely matches for performance reasons given that most customer records in a file do not match. We also often match customer records using multiple match keys/match criteria when one match key will not capture enough desired matches. Spectrum's Advanced Match Module Intraflow Match stage has a limitation when you want to match records within a Single Data Source using the combination of Multiple Match Keys. The limitation is that you can only use only one  Match Key per Intraflow Match stage and Spectrum does NOT provide an easy way to do a union of all matching records from each match stage that uses a different match key. The work around for Matching Within a Single Data Source Using Multiple Match Keys is to create a Intra Batch Matching job with the Match Key Generator stages and Intraflow Match stages that uses more than one match key to effectively match on multiple separate criteria and then run an external application (e.g. Python or Java) through a Spectrum Process flow do a union of all matching records from each match stage that use a different Match Key.

     

    Attached is a SCV Advanced Topic 'How To' Short Training Video on how to implement this functionality.



    ------------------------------
    Jeremy Peters: GISP, CIMP - Master Data Management & Data Quality
    Solutions Architect / Principal Consultant/ Distinguished Engineer
    Pitney Bowes Customer Information Management and Location Intelligence
    350 Jordan Rd | Troy, NY 12180 USA
    jeremy.peters@pb.com
    https://www.linkedin.com/in/jeremypeters1/
    ------------------------------