chergh - Thanks, but like I said this is not a dB, its using a database (the only way I could think how to) reaching a required result; I know how to normalise data etc - however this is one occasion were, well as you can all tell there seems to be no easy way of reaching a swift conclusion.
gemma-the-husky - thanks kind of what I've done in a flipped way, I've used union to pull all the files together - doing this by id would take ages as there are a couple of thousand people in the file.
Thanks
H