In our database we have since long a table with file paths. A file path typically has a length of up to 300 characters. Each path is analyzed before storing and a single, unique record is made per volume, folder or file. The total file path is an unstored recursive calculation. By the way, we had this working in FMP 6 too!. Sometimes the analyzing script is mistaken and an existing folder path record is made a second time. Up to FMP11 it was possible to do a search for duplicates in the unstored calculation of the file path field. Since FMP11 this doesn't work anymore. Apparently the number of characters searched is limited to around 200. Is this documented?
Is there a work around?
We tried several things, one of which is calculating a simple hash number. The resulting hash codes were never longer than 400 digits. The FileMaker Help tells us about numbers: "FileMaker Pro indexes the first 400 significant digits (numbers, decimal points, or signs) of the field, ignoring letters and other symbols." So I would think that it is possible to search for duplicate hash numbers, but this doesn't result in correct results either. Apparently FMP11 uses far less digits then 400 for the search.
The attached file contains 2135 file paths both as text and as a unstored calculation and as a hash.