Fuzzy matching on financial data is often based on data entry problems (intentional or accidental). Primarily numeric based data like invoice numbers, purchase orders, contract numbers and even phone numbers are often affected by data transpositions, typing one letter for a digit or swapping digits (10 instead 01).
The SIMILAR() function offers an opportunity to find similar values by first harmonizing the data. By replacing alpha characters with their numeric equivalents (1 for I, 0 for O, etc.), this function allows fuzzy matches on financially coded information to be easily made. Worst case, a few false positives may be identified, but few, if any, likely matches will be missed.