The xargs command in Linux
xargs reads items from standard input and uses them as arguments to another command. It bridges a gap in the pipe model: pipes send data to a command's input, but many commands (rm, cp, mkdir) take their operands as arguments instead. xargs converts one into the other, and can parallelize the work with -P.
How xargs works
There is a hard limit on how long a command line may be (ARG_MAX, a couple of megabytes). xargs exists at that boundary: it reads items from stdin and packs them into as few command invocations as fit under the limit — turning "run rm on these 200,000 files" from an "Argument list too long" error into a handful of maximally full executions. It converts a stream into argument lists, which is the missing adapter between pipes and commands that take operands as arguments.
Its historical weakness is tokenization: by default xargs splits on any whitespace and interprets quotes, so filenames with spaces shatter. The modern fix is a different contract — NUL-separated input, produced by find -print0 and consumed with -0, unambiguous because NUL is the one byte a filename cannot contain. Add -P and the same machinery becomes a parallel executor: same batching, N workers. find | xargs -0 -P is a small distributed system in twenty characters.
Syntax
SOMETHING | xargs [OPTIONS] [COMMAND]Common options
| Option | What it does |
|---|---|
-I {} | Placeholder mode: run the command once per item, with {} replaced. |
-n N | Pass at most N items per command invocation. |
-P N | Run up to N invocations in parallel. |
-0 | Expect NUL-separated input (pairs with find -print0; safe for spaces in names). |
-r | Don't run at all if input is empty (GNU). |
-p | Prompt before each invocation. |
How to use xargs: examples
$ find . -name "*.tmp" | xargs rmDelete every .tmp file found. find lists them; xargs feeds them to rm as arguments.
$ find . -name "*.log" -print0 | xargs -0 rmThe safe version: NUL separators survive spaces and newlines in filenames. Make this the habit.
$ cat urls.txt | xargs -n 1 -P 4 wget -qDownload a list of URLs, four in parallel.
$ ls *.jpg | xargs -I {} cp {} backup/Placeholder mode: run cp once per file, inserting the name where {} appears.
$ echo "dir1 dir2 dir3" | xargs mkdirCreate several directories from a generated list.
Real-world use cases for xargs
Mass operations on find results
Compressing hundreds of old logs: find /var/log -name "*.log" -mtime +30 -print0 | xargs -0 gzip. The -print0/-0 pair makes filenames with spaces safe, and xargs batches them into few gzip invocations instead of hundreds.
A minimal parallel job runner
Processing a list of items with 8 workers: cat items.txt | xargs -n1 -P8 ./process.sh. For embarrassingly parallel work — image conversion, URL checking, per-host commands — this replaces a queue system in one flag.
Pro tips and common mistakes
- Preview before executing: insert echo — … | xargs echo rm shows the exact commands that would run.
- Make find|xargs pairs NUL-safe by habit: -print0 on find, -0 on xargs. Whitespace in filenames stops being a threat.
- GNU xargs -r skips running entirely on empty input — otherwise some commands run once with no arguments.
- -I {} enables argument placement anywhere (and implies one-per-invocation): … | xargs -I {} mv {} done/{}.
Frequently asked questions about xargs
Why use xargs instead of find -exec?
find -exec cmd {} \; launches one process per file; xargs batches many files per invocation, which is dramatically faster on thousands of files. (find -exec cmd {} + batches too, and is a fine alternative.)
How do I handle filenames with spaces?
Use the NUL-separated pair: find … -print0 | xargs -0 …. Plain xargs splits on whitespace and will mangle such names.
Can xargs run things in parallel?
Yes — -P N runs up to N invocations concurrently, turning xargs into a minimal parallel job runner: … | xargs -P 8 -n 1 process.
Related commands
Delete files and directories — and understand -r, -f and why there is no trash bin.
grepFind lines matching a pattern in files or piped input — the workhorse of text search.
findLocate files anywhere in a directory tree by name, type, size, date — and act on them.