Manual Resolution
Provides an overview of steps required to manually resolve partial matches in the partial confidence bound.
Manual Resolution Overview
In order to enable manual resolution of partial matches where human intervention is necessary, a column labeled resolve can be added to a list of resolution result to override the existing result with the status of no_match when a new identity should be created ormatch when a record should be matched to an existing identity.
Manual Resolution Process
1. Create the resolve column
Add a column labeled resolve to the partial_match.csv file. Although it makes most sense to use the partial match file, this process can be applied to any conformant CSV results file, including the created.csv file or the matched.csv file.
2. Label the rows to resolve
In the resolve column in the partial_match.csv file add a T to the row that will be manually edited.
3. Update the incorrect fields
Manually update the fields in the row so they match the fields that start with existing. This will fully update the manually resolved record. Save and rename the partial_match.csv file as:
~/file_location/fusd_02_10_2020_resolved.csv4. Add the associated schema
Copy the existing schema for the agency and rename it in the following manner.
cp ~/file_location/fusd_02_10_2020_resolved.json ~/adapter-location/deposit5. Start the adapter
Similar, to the instructions in the Adapter setup section, copy the new schema to the ./deposit file. Enter the root directory and executing the following script to start the adapter:
Running this script will start the executable locally. If started successfully, the terminal should display the following start screen:
The adapter is now running locally and waiting for an import file to be dropped in the ./deposit folder
6. Depositing the manual resolution file
Place the manual resolution file in the ./deposit folder located in the root directory:
7. Conclusion
At this point all manually resolved records will have been resolved and matched against existing records. Because they were manually adjusted to match the existing records, they will create a new ~resolved_matched.csv file located in the ./results directory:
Export into Agency System of Record
The final step in the process is to load the universalIdentifier created and/or match on into the agency's system of record. All CSV's are prepared for agency import into its system of record and the new universalIdentifiercan be stored alongside the agency-specific reference identifier.
Last updated
Was this helpful?