*sorry in the title> "too" means "two" my bad.. HI people! I wrote a code that deleting a file from foldor B if it dont exist on A: for /f "delims=" %%a in ('dir /b "D:\B"') do ( if not exist "D:\A\%%a" del /Q "D:\B\%%a" ) has you can see the %%a is the file name. To make it more clear, lets say that in foldor A we got: 1.txt 2.txt And in foldor B we got: 1.txt 2.txt 3.txt The file 3.txt will be deleted. NOW Here is my problem: The code works fine on a simple directory but if i try it on foldor with subdirectories it wont do a thing.. How can i make it work on the subdirectories as well? *I am very sorry about my english i don't use to write so much...

i can't write winXP batch script because my XP system is dead, but this might be the logic structure: 'path1 is the path you are using as the standard path1="c:\dir1\dir2" 'path2 is the one you want to "clean up" path2="d:\dir1\dir2" cd path2 call foo exit (that is:) goto gtfo foo: for x in (dir /b /a:d) do ( pn=p1 p1=path1+"\"+x cd x call foo ) for y in (dir /b *.*) do ( if not(exist) p1+"\"+y del y end if ) cd .. 'much apologies, i KNEW that was too simple! 'the path-string manipulation much more gnarley: 'this is the gnarly part i didn't count on... 'base dir. of C for example: C:\uuu\bbb 'base dir of F (whatever, network drv) is: F:\y\z 'whatever locn of F, need to snip off "F:\y\z" 'from curr. locn ("CD") and repl 'with "C:\uuu\bbb" base path to test 'for 'existance. 'via a "dir /a:d /b" command or "if exist" test. 'get curr. fullpath on slave: can you do this? set p=("cd") 'snip off the base 'is there such a thing as len()? set n=len(path2) 'is such thing as mid (or substr, etc) set p=mid(p,n+1,100) 'now stick it onto the master basepath set p1=path1+p 'you can see why i am currently limited, 'but others in forum will know how to do 'this snip/cut/paste stuff. p1=pn if not (exist) p1+"\"+x rd x end if return :gtfo i'm not sure if this would work, i don't have the equipment to try it or the syntax knowledge to write it. the idea is that "foo" keeps calling itself until it reaches the periphery of the tree structure, then starts conditionally deleteing all the files not on drive C of the same name and path. the "rd" (rmdir, remove directory) is not conditional cause i couldn't think of a quick and easy way to see if it is empty, but if it's not empty, the "rd" will fail anyway, so no harm done. my logic might be horse-dukey, so if you manage to translate it into working code, be careful about testing. ps: my mistake, forgot to put the path-strings in the code. (p1, pn)