Here's a way to do it. The first section just sets up some dummy data. The second part results in a 2D array where each row is an array of matched strings based on the first 8 characters of the string. The last part indexes this 2D array and truncates each matched string array to eliminate empty string elements. Inside the blank sequence structure is where you would perform the action(s) you want on each matched string array.
This is not very fast. Depending on the randomness of the dates, it could take up to 1 minute to process (using a P4 CPU). The more dates that match, the less time to process.
Dave.
==============================================
