Some data matching tasks don’t succumb to a single matching condition. In these cases, it’s common to use an ordered list of rules to process the data one at a time, reducing the unmatched items gradually as each rule is applied.

An example cascade for cash flows might be:

  1. Match entries with identical references on both sides
  2. Match entries with exact numeric value
  3. Match entries falling on same value date
  4. Match totals per day

Depending on your background and area these may seem reasonable or not. Sometimes a balanced total for the month is good enough, sometimes only exact multi-column criteria is appropriate.

Of course, a single outcome is also not necessarily desired. An exact match might trigger the release of an order, while approximate matches get flagged for manual confirmation and omissions are ignored for a grace period before being queued for a support function. Many flavours and combinations are possible.