finding renamed files

[Alex Tweedly]

David Glasgow wrote:
> I am looking for a quick and dirty method for walking a directory and 
> finding files that have been renamed by the user.  I don't need to 
> find them all, just as many as possible.  The folders are likely to 
> originally contain matching stems and progressive numbers pjf017.jpg, 
> pjf018 .jpg, pfj019 .jpg etc. etc, with the user renamed files 
> standing out completely arbitrarily by not following the pattern .
>
> At the moment I do this using the eyeball test, which is remarkably 
> quick and efficient but very very very boring because there are often 
> thousands of files to scan.  One approach I thought of is to 
> progressively filter the folders' contents by nibbling a character off 
> the end of the first filename.  If it is completely unique (and 
> possibly therefore renamed), nothing will happen.  However if  9 other 
> files disappear, it was a name representative of progressive pattern.  
> Nibble another character, and so on until it is gone, and any 
> filenames left over didn't fit the dominant pattern in the folder.  
> Yes?  No? .  Any other suggestions?

I'd try to exploit the fact that such file names usually follow a 
pattern of
   <prefix><number>.<suffix>


So the *really* quick and dirty method uses that, and also simply 
assumes that the <number> part will be 4 digits (true for most cameras) ...
(beware typos - not tested)

set the itemDel to "."
repeat for each line f in the files
   if char -4 to -1 of (item 1 to -2 of f) is not a number then put f & 
CR after theChangedList
end repeat


Obviously, there are less quick, and less dirty, variants that actually 
check the prefix is used multiple times - but for me, this would be good 
enough.




-- 

Alex Tweedly      mailto:alex at tweedly.net      www.tweedly.net