Please read my contribution to Topic “How to retrieve a lot of documents …” in this forum.
To unload nonxml documents you would use an “ino:id” CURSOR, as described there. Additionally, you may want to set up another CURSOR with the following query:
Then, when looping over the 1st (ino:id) CURSOR and do GETs on each ino:id , you may want to write out the (nonxml-)documents one by one and use the ino:docname from the 2nd CURSOR as filename, writing the docs all to the same directory.
To load the documents again, you would loop over the files in the directory, doing HTTP PUT for each file to the new destination. It should be easy to write a perl script / java- /c-program to do that.