I could discuss some of the search and replace scripts I have that use Ruby’s regular expression capabilties, but in the interests of simplicity, you might want to try Barebones TextWrangler: http://www.barebones.com/products/TextWrangler/textwranglerpower.html
It supports multi-file search which would address your needs in a relatively interactive user interaction.
It doesn’t do RTF, but you could search on the character escapes and replace accordingly.
Another choice would be Text-Edit Plus: http://www.tex-edit.com/
Which will open RTF and is eminently Applescriptable. Downside is that it doesn’t do RTF Unicode escapes, which I saw that your post contained:
\'e5 to \uc0\u257 \
\'c4 to \uc0\u256 \
\'e0 to \uc0\u7693 \
\'e2 to \uc0\u7692 \
It does the multiple find and replace for RFT and also does multi documents. It also has some other features that are good for cleaning files of characters ect.
It is designed for writing code like BBedit and Textmate, but both those app did not work for this task here.
As a work flow it was very easy to export the file out of devonthink, process it through jedit x in a flash and import it straight back into the archive.
A cool feature is that you can export the find and replace set out of Jedit X as txt files and adjust them and import the back in.
This made it easy for me to have different set to work with.
My final note on this thread is the happy discovery that with jedit X you can drag a file out of devonthink - not export- into jedit X use it’s text processing tools not available in devonthink and save the changes directly in your devonthink archive/database.