Infer Value

 

Purpose

Infer a value for attribute A using a known relationship between attribute A and attribute B.

 

Field Description

  • Mapping data source – Choose a reference data source.
  • Compare the input value from – Choose an attribute from the input data source pick list to compare
  • To look-up values from – Choose an attribute from reference data source picklist to fetch values
  • Using the match method of – Select Exact match, Fuzzy match, Begins with, End with or Contains.
  • Output value from – Choose an attribute from the reference data source to write.
  • If more than one matches are found – Choose from the select picklist to write single value or all matching values ( up to 100). For details on using the “Write single value using counter” option, refer to Assigning Leads using Even Distribution or “Round Robin” Style
  • Resolve the conflict using priority in – Choose the priority from picklist. This requires the reference data source has a priority column for conflict resolution.
  • Write output to – Choose an attribute in the input data source to write the inferred value.
  • Advanced configuration description:
    • Apply filter to reference data – Use this option to filter reference data.
    • Allow match on blanks [] for: – select to enable blank matching.

 

Tips

  • This task requires the use of a Reference Data Source that contains the relationships between attribute A and attribute B. For example, the Openprise reference Reference – States, Provinces, Regions contains state and country relationships. This can be used to infer missing country information if the state information is available.
  • Matching can be done on multiple attributes, eg. infer a value for attribute A by matching on attributes B1, B2, B3, … etc.
  • You can output more than one attribute, and you can also output custom text.
  • It is best not to overwrite the existing value in Attribute A, so instead use a filter for attribute A = Has no value to infer a value for attribute A only if it is empty, or write the results to a new attribute.

 

Examples

  1. Infer missing country attribute if there is recognizable state value.
    • For example, if State = “California”, then infer Country = “United States”.
    • Use the Openprise reference Reference – States, Provinces, Regions.
  2. Infer missing postal code from a known city and state data.
    • For example, if City = “Redwood City” and State = “California”, then infer Postal Code = “94065”.
    • Use the Openprise reference Reference – US ZIP Codes.

 

Advanced Usage

Openprise has a special value that can be used in reference tables along with the Infer value task to perform wildcard matching. The value is “OP_MATCH_ANY”. The following example illustrates how this feature is used.

Example: Suppose I want to assign a score using State and City. When State = NY, and City = New York, I want to assign a score of 2. For all other cities where State = NY, I want to assign a score of 1.

  1. Begin by creating a Data Source reference table, with the contents as shown below. I’ve named my reference Data Source “Infer Value OP_MATCH_ANY”. Remember to import the data after you create the Data Source definition.
  2. Infer
  3. Note the use of “OP_MATCH_ANY”, which is a special value that allows wildcard matching.
  4. Then, I create a Task that uses the Infer value task template and looks like this:
    Infer

  5. NOTE:  I have configured the Advanced Configuration section to allow matching on blanks for “Matching Criteria 2” which is the “City” match.  This is to ensure that a score of 1 is set for a City that is blank but State that is NY. Without that checkbox, we will only assign a score of 1 if the State is NY and City has any value other than New York or blank.
  6.  

  7. The result of running the above Task is shown below. You can see that Openprise will prioritize a full match of both criteria over a wildcard match, thus achieving the results I desired.
  8. Infer

Support Contacts

If you have any additional questions, please feel free to contact us at support@openprisetech.com.

 

Leave a comment