How do I remove files downloaded to a directory that are now mixed in with other files without deleting the other in Linux

Question

I'm a Linux user. I accidentally downloaded a bunch of files into the wrong directory. They're now mixed in with my other files. I then created a new subdirectory and downloaded all the files again, this time into the subdirectory. What is the best way to remove all of the files I accidentally downloaded to the directory without accidentally deleting any of my existing files. I'm new to Linux and need some help. I suppose it could be down by date, or it could be done by saying for each file in the subdirectory, delete a file in the directory with the same name. Thanks in advance! -- Larry

Answer 1

If the file modification times are consistent, then you can use a variant on touch and find to locate the files to be removed. Suppose you downloaded the files just after 12:34 today (2009-09-08):

touch 20090908123400 empty.file
find . -newer empty.file -print

This will list the files. When you're sure you've got the right files, you can revise the find command to:

find . -newer empty.file -print0 | xargs -0 rm -f

Using the '-print0' and '-0' options means that the pipeline will handle file names with spaces and all other special characters correctly. You could also use -older and a second empty file with a different time stamp to frame the time.

If the downloads preserved the modification times of the files on the remote system, you are probably hosed. However, that is rather unlikely to be a problem.

Answer 2

Actually this is very easy to do. cd into the directory that you accidentally downloaded the files and run the following code, replacing subdirectory with the name of the subdirectory that contains the files you downloaded a second time.

rm `ls subdirectory`

What this does is list all of the contents in subdirectory and this list is passed to rm which Re Moves any files with the same name in the current directory.

Answer 3

To expand on Pynt's conceptually correct solution, you can handle files with spaces in their names by using the null-separator options of find and xargs:

Suppose your downloads folder is ~/Downloads/, which contains the files file 4 and file 5. You actually wanted to download a set of files (file 1, file 2, and file 3) to a subdirectory ~/Downloads/subdir/. You accidentally downloaded those files to ~/Downloads/, then you downloaded them correctly to ~/Downloads/subdir/ as well. So, your current situation looks like this:

$ tree ~/Downloads/
.
|-- file 1
|-- file 2
|-- file 3
|-- file 4
|-- file 5
`-- subdir
    |-- file 1
    |-- file 2
    `-- file 3

To delete the right files, try this:

$ cd ~/Downloads
$ { pushd subdir; find -mindepth 1 -maxdepth 1 -print0; popd; } |\
  xargs --null -I'{}' trash '"{}"

That code will not delete the files. It will send them to the trash instead, as long as you have installed the trash-cli package. (See http://superuser.com/questions/32355/undo-linuxs-rm/32480#32480 for more info on the trash command, which I highly recommend.) If you'd rather live dangerously, replace trash with rm. If you want to live even less dangerously, replace trash with echo trash. That will simply print the commands that will delete your files.

Anyway, here's the explanation, since you should never trust code you don't understand. The first line simply moves into the Downloads directory. The second line finds and prints the names of all the files in subdir. pushd and popd are just like cd, only with a back button. (Look up stacks if you wonder where "push" and "pop" come from.) The options to find tell it to only list things in subdir and not, for example, subdir/another subdir/third subdir. They also tell it not to list ., the name for the current directory itself. You don't want to remove that. Lastly, the -print0 option tells find to print the files it finds, but separate them with null characters instead of whitespace. This is how we deal with spaces.

Likewise, in the third line, the --null option to xargs tells it to read the filenames from find with null characters separating them. The rest of the third line constructs the command that will delete your files.

The lesson in general is that you need to process files that might have spaces, you generally need to use a command that involves find -print0 | xargs --null.

Answer 4

Simple bash solution

cd subdir
for i in *
do
    rm "../$i"
done

I haven't tested it, so you'd better test it before you run it for real :)

Answer 5

Pynt and Peltier have this basic idea, but only Peltier's will handle files with whitespaces.

This should work:

cd path-to-subdirectory
set -- *
cd -
ls -ld "$@" | more    # Make sure the list is what you expect to remove, then:
rm -rf "$@"

I show using "ls" to confirm the file list before removal, for sanity.

But you could also move them to a temporary place (including a ZIP archive) if you're concerned about deleting the wrong stuff. Just replace the "rm" above with either of the following, then if you trust what happened, remove the ZIP or temporary directory:

mkdir /tmp/to-nuke ; mv -vi "$@" /tmp/to-nuke
zip -0rm backup.zip "$@"