Occasionally I’m away from the bash shell for too long and I forget some things. This is just a quick reminder to myself regarding the syntax and usefulness of xargs.

Why?

xargs takes a file list as input and performs some arbitrary operation on it. This is obviously useful for scripting and bulk file operations. Why use xargs? Well, quite frankly, it’s faster than using find -exec because it splits the file list into sublists and calls the command once for every sublist instead of calling the command once for every file. Also, because it uses sublists you never run into problems when trying to run commands with very large argument lists (it is actually possible to run a command and have bash reply that the maximum line length is exceeded). Below is example usage.

Example: fixing permissions

Using xargs we can easily correct permissions recursively. For instance, let’s improve our web security by locking down write privileges on our directories and files to only the owner:

find . -type d -print0 | xargs -0 -I {} chmod 755 {}
find . -type f -print0 | xargs -0 -I {} chmod 644 {}

Here’s what’s happening:

• we use find and -type d to locate all directories under the current directory
• we use -print0 so find will print the directories and separate the directory names by null characters instead of newlines; delimiting this way improves xargs ability to handle special characters in file names
• then we pipe this list to xargs and use -0 to tell xargs the list is null-character delimited
• we also use -I {} to tell xargs to replace any occurrence of {} with each filename as it runs the following command
• ultimately the command chmod 755 {} tells xargs to chmod 755 each directory

Of course we then do the same by running chmod 644 on all files under the current directory.

Efficiency

Using xargs this way is faster than running these equivalent find commands:

find . -type d -exec chmod 755 '{}' \;
find . -type f -exec chmod 644 '{}' \;

Using find this way will run chmod individually for each directory and file, which is less efficient. The general rule of thumb is to always opt for xargs.

You can use xargs for several other things, but this example is my most frequent use-case. Happy hacking!