I have a filemaker database with a field with a very large amount of text per record. I would like to analyze the whole field, for all records, for number of " word " occurrences for all text. Essentially what a keyword density tool does but for filemaker. I know that there are several calculations, and procedures for finding occurrences, but none that I know of, will work without specifying the text that you are looking for. Please help!

none that I know of, will work without specifying the text that you are looking for

If by that you mean that the keywords should be taken from the text itself, you will need a custom function (requires the Developer version) or a scripted process to go over the text a word at a time.

Note also that it's not easy to find the number of times a word occurs in text as a word, i.e. ignore "applesauce" when counting "apple".

well in that case do you know of any keyword density tools where you can upload a giant csv file or something. The situation is basically that we have a large database with a large text field for each record, and we need to group the database by the keywords that occur most in the text field (without knowing what those keywords are). I haven't been able to find any, though I'm sure they exist (it's a mess trying to find something on google because of all of the SEO that the keyword density tool companies do :(

I am afraid not. I would still do this in Filemaker - though I'd hate to do it in version 7.

How large is the text in the field?

What would the output of such a keyword density tool look like?

Would it be a delimited structure with each row a ?

I guess it depends on what you want to do with it, once you have it. For any serious use it should be records in a related table.

There is a free program called word counter http://www.supermagnus.com/mac/Word_Counter/

It's free, and is cool, but cant process a file nearly that large. If there were something like this that was web-based and enabled for large files that would work. I would like to make it table related in FM, but without knowing the keywords first, it's a challenge