I have a DT3 db with 419k items. Many are duplicates from incorrectly importing or allowing duplication of items on import (which I have since fixed). Most are files with identical creation dates and strings in the name that begin with the same initial characters. I have not found a way to create a smart rule based upon matching creation date and size (but that has been my manual sort and click solution). Another possibility could be an Applescript that recursively sorts the database and moves files meeting those two criteria into the trash or a holding file. Yet another approach might be a regex search for the shortest string as the name and other longer names whose creation dates and size match being the result. I know enough about DT, scripts, CLI, and regex to know that it can probably be done, but… 1) find files with identical creation dates, 2) verify matching size, 3) verify matching ‘begins with…’, 4) delete dupes, i.e., all the files except that with the shortest name.
Smart rules can search for files matching the specified critieria but it’s not possible to recognize duplicates based on your own conditions. The only possibility would be to use a script.