I am working with a flat file from a supplier and need to get it into a reliable CSV format.

I am using filemaker to port it to this format.

One of the issues is that the description field they provide is messy. It has many characters that do not play nice.

I am stripping a few using a script to find and replace things such as # signs, commas, semicolons, and other characters to get a clean text.

One issue is that I need to remove the " in some fields - but I cannot find a way to do a find and replace on "

How can I clean the fields that have these - it needs to be something scriptable since this process will be ongoing.

Rather than Find and Replace, I'd have a calc field that makes use of Filter ( ) and Substitute ( ). Have a look at some CFs on Dunning, like CF for Cleaning Text.

find and replace things such as # signs, commas, semicolons

These characters are perfectly legitimate and you should have no problems getting a valid .csv file when exporting a field containing them.

Perhaps your target application cannot handle a standard .csv file.