Performing an External Match

You can use a model with scope as external to match records from a file with the records in the tenant.

This task explains how to select the input file that contains entities for the purpose of matching. It also provides information on selecting an existing mapping or creating a new one.
  1. Click on the link shared by Reltio to open the ML Match application. When you log in and select your tenant for matching, the following screen appears:

  2. Click ADD NEW and select the entity type for which you want to run the match. Only entity types that have an associated model appear for selection. For example, if you select the HCP entity type, it indicates that the input file contains a list of HCPs and the model associated with the HCP entity type will be used for matching.
    You are prompted to upload the file that contains records that need to be matched. Select the location of the file. You can choose from any of the following:
    • Amazon S3
    • Google Cloud Storage
    • Local file system
    Note: You must ensure that the column headers in the input file do not start with a number and contain only alphanumeric characters.
    To select a file from the local file system, click BROWSE FILE and select the input file. The following image displays that a file from the local file system is selected:

  3. After you select the file, click CONTINUE.

    You are prompted to select an existing mapping. However, as you are creating the flow for the first time, no existing mappings appear.

  4. To create a new mapping, click CREATE NEW MAPPING in the Select mapping window.

  5. After providing a name for the mapping, click CONTINUE.

    By default, the first column is selected as the Record Identifier. When you download match results, the record identifier allows you to identify the input records that matched with the records in the tenant.

    You can change the Record Identifier if required to any other column that contains unique values across your dataset.

  6. To change it, click on More Options next to the relevant column and select Set column as record identifier.

    The following image shows that the Id column is set as the record identifier:

  7. Click on a column header and map the required attribute from the panel that appears on the right. You must map all attributes with the columns in the input file because the displayed attributes have been used to build model.

    The following image shows that attributes used to build the model are mapped to the relevant columns:

  8. Click RUN MATCH.
    The following image appears indicating that the matching process is initiated:

  9. Click OK.
    The following image appears indicating that the job is in the In progress state:

    When you hover the mouse next to the job status and click the More Options button, you can download or delete the job definition. If you download the match results, a ZIP file is downloaded that contains matches.csv or non-matches.csv files based on the match results. The matches.csv also contains a column called relevance that contains the relevance score of the matched records.
    Note: You can delete an external match model flow only if it is in an error state.