We have a legacy solution (started in 1993). It is currently in FM6. In this solution we have several file with extensive repeating fields which were necessary to perform complex searches. Most of the data was entered via pre-OSX Macs. we are now only OSX Mac and PCs. The problem is that there are charcters in some of the fields which are not recognized by the PC (apostrophy). The field contains very long chemical names with many of these (known as primes in chemical speak.) This seems to be causing the file to crash on import or export. There were also corrupted layouts which we have eliminated. Eventually, we will convert these to related daughter records in FM8, but for now they must remain in FM6 as repeats. We cannot export to a non FileMaker format and retain the repeats as was recommended by a professional. How do we clean this up?

One thing you could do is use Replace with a calculated result to change the apostrophes to straight quotes.

I don't get this part:

We cannot export to a non FileMaker format and retain the repeats as was recommended by a professional.

I don't know what was recommended by a professional, but you certainly can. Use the merge format, for example, and be careful not to remove the invisible group separator character (ASCII 29) while cleaning up the file.

New Twist: We recreated the file from scratch. Imported from the original file (on a Mac to avoid crashes) by exporting as a merge file, opening on a Mac in FM 8.5, running a text formatting script on the data to assure no bad characters), re-exported as a merge and imported into the new FM6 file. Entered a new record which uses 34 of the 250 repeats (on a PC). At the 34th repeat, FM crashes. We went to other records of that or greater length and the 34th repeat crashes from any record.

Any ideas why this repeat(or anything beyond it)would be corrupted on a brand new file? Created on a PC, run on a PC and data entered on a PC. crashes on a PC, but not a Mac.