How do I configure match rules to ignore certain special characters?

Colin Parker

Updated January 07, 2021 13:44

Question

Need to set up match rules so that certain characters are ignored, possibly requiring that the data not be changed. How can I configure that?

Answer

You can achieve this in Reltio by using the String Replacement Cleanser with regex. This cleanser replaces predefined strings part of attribute values with more meaningful strings to improve matching, ex: replace “Mem Hosp” with Memorial Hospital. The string is replaced when the match document is created, and the updated values are stored in the match document only for tokens and/or comparator use. It will not change the values in the attribute.

Alternatively, if you would like to change the data to remove the special characters for better matching, you can use the String Function Cleanser (scrub parameter).

The following use case helps you to understand how the two cleansers work one after the other:

Sample Address

{
 "Address1": "12021 Wilmington (Ave) 1st$ Fl",
 "City": "Los Angeles",
 "State": "CA",
 "Country": "United States",
 "Pincode": "90059"
}

At first, the above address is cleansed for any special character present in the address. This address is then verified with geo-location.

The Reltio StringFunctionCleanser cleanses the address for the special characters and then the AddressCleanser cleanses the entire address.

Expected Address Output after Applying the String Cleanser

The above sample address gives the following output after applying the StringFunctionCleanser to scrub the special characters:

{
 "Address1": "12021 Wilmington Ave 1st Fl",
 "City": "Los Angeles",
 "State": "CA",
 "Country": "United States",
 "Pincode": "90059"
}

Reltio recommends you to use different resultingValuesSourceTypeURI (crosswalk) for StringFunctionCleanser and AddressCleanser in cleanseconfig

cleanseConfig": {
    "infos": [
      {
        "uri": "configuration/entityTypes/HCP1/cleanse/infos/StringFunctionCleanser",
        "useInCleansing": true,
        "sequence": [
          {
            "chain": [
              {
                "cleanseFunction": "StringFunctionCleanser",
                "resultingValuesSourceTypeUri": "configuration/sources/StringCleanser",
                "proceedOnSuccess": true,
                "proceedOnFailure": true,
                "mapping": {
                  "inputMapping": []
                },
                "params": {
                  "scrub": "(#%!~)"
                }
              }
            ]
          }
        ]
      },

References

https://docs.reltio.com/matchmergeconfig/replacecleanser.html?hl=regex

https://docs.reltio.com/datacleansenew/cleanselibstringnew.html?hl=stringfunctioncleanser