2014 PINES Patron Database Cleanup

Date: 
2014/09/17

Overview

In August/September 2014, GPLS contracted with Emerald Data Networks and Unique Management Services to provide us with address updates and to identify duplicate patron accounts for the PINES system administration staff to process.  The PINES staff processed the address updates on 8/28 and 8/29/2014.  The de-duplication process began in mid-September and is expected to take several weeks of nightly batch processing to complete.  Patron accounts identified for automated de-duplication will be updated with alert messages indicating that they are a candidate for automated merging to another account.

Merge Criteria

Here are the criteria the PINES staff provided to Unique Management Services to identify the "lead" account (the account which will still be considered "active" after the merges are completed):

  • "lead account" = the account that we want duplicates to merge onto
  • "duplicate account" = accounts that duplicate the lead account

Given a set of duplicated accounts, the lead account is identified by the following:

  • the account expiration date is the newest OR
  • the account is barred OR
  • the account is in collections with UMS

Accounts in collections with UMS:

  • if one or more duplicated account is NOT in collections, it should be merged with the collections account
  • any duplicates that are also in collections should NOT be merged

With those parameters in place, Unique Management Services has returned a file to us that contains a "match score".  Here is an explanation of those match scores:

match score Description
1 Patron first name, middle name, last name, suffix and date of birth match exactly. (e.g. "Jonathan", "Livingston", "Seagull", "Jr.", "1970-01-01")
2 Patron first name, last name, suffix, and date of birth match exactly. (e.g. "Jonathan", "Seagull", "Jr.", "1970-01-01")
3 First 5 characters of patron first name, last name, suffix, and date of birth match exactly. (e.g. "Jonat", "Seagull", "Jr.", "1970-01-01")
4 First 2 characters of patron first name, last name, suffix, and date of birth match exactly. (e.g. "Jo", "Seagull", "Jr.", "1970-01-01")
5 Patron first name, middle name, and last name match exactly (e.g. "Jonathan", "Livingston", "Seagull")

After careful review by Unique Management Services staff, Emerald Data Networks staff, and the GPLS PINES Staff, we automatically merged accounts with a match score of 1 or 2.  Soon the PINES staff will distribute lists of potential matches that require manual review.