Cut, Sed, Awk, and Xargs

Objectives

By the end of this chapter, you should be able to:

• Explain what cut does and when to reach for it
• Explain what sed does and when to reach for it
• Explain what awk does and when to reach for it
• Explain what xargs does and when to reach for it

💡 Why this matters: Every command in this lesson exists to reshape or process text without opening an editor — slicing columns out of a file, replacing text across thousands of lines, or running one command across a whole batch of files. This is what “power user” actually looks like in Terminal.

cut

cut pulls specific pieces out of each line of text. Given a file languages.txt:

Java,James
Ruby,Matz
Lisp,John
Bash,Brian
Self,David

You can cut by character position — the -c flag grabs a range of characters from each line:

cut -c 1-4 languages.txt

Java
Ruby
Lisp
Bash
Self

That works here, but only because every language name happens to be four characters. A more reliable approach splits on a delimiter instead — -d sets the delimiter, -f picks which field to keep:

cut -d, -f2 languages.txt

James
Matz
John
Brian
David

-f1 would give you the first field (the language names) instead. Delimiter-based cutting works regardless of how long each piece is, which position-based cutting can’t guarantee.

cut also chains naturally with what you already know:

cut -d, -f2 languages.txt | sort | head -n 2

That pulls out just the names, sorts them, and keeps the first two.

sed

sed (Stream EDitor) is a much deeper tool than this lesson can fully cover, but its most common use is find-and-replace. To swap every comma in languages.txt for a colon:

sed 's/,/:/g' languages.txt

Breaking that down: s means substitute, , is what to find, : is what to replace it with, and g means do it for every match on the line, not just the first.

Java:James
Ruby:Matz
Lisp:John
Bash:Brian
Self:David

Run cat languages.txt afterward and the file is unchanged — by default, sed only prints the result, it doesn’t edit the file. To edit in place, add -i, but the exact syntax depends on your system:

# macOS (BSD sed) — needs an explicit backup suffix, even if empty
sed -i '' 's/,/:/g' languages.txt

# Linux (GNU sed) — -i works on its own
sed -i 's/,/:/g' languages.txt

cat languages.txt now shows the replacement actually saved. This is genuinely just the surface of sed — keep pushing on it as you run into real find-and-replace problems.

awk

awk is powerful enough to be considered its own small language for text processing. The simplest form just prints a file: awk '{print}' languages.txt. Where it gets useful is printing specific fields. With -F setting the delimiter, $1 refers to the first field:

awk -F ':' '{print $1}' languages.txt

awk’s default delimiter is whitespace, so no -F is needed for space-separated input. Try it on your command history:

history | awk '{print $2}'

That prints just the command name from each history entry, dropping the line number and everything else.

awk also handles row-and-column data. Run df -h (short for “disk free, human-readable”) on its own first to see the raw table, then pull one specific value out of it:

df -h | awk 'FNR == 2 {print $4}'

FNR is the current line number — this grabs the 4th column of the 2nd row, which is typically your available disk space.

xargs

xargs takes a list of inputs — usually piped in from find or ls — and runs a command once for each one. That turns a single-file command into a batch operation:

Try It

All four commands process real files, so try these yourself:

1. Recreate languages.txt and use cut two different ways — by character position and by delimiter — to see them agree on this particular file.
2. Use sed to replace a word across a text file of your own, first without -i to preview it, then with -i to actually save it.
3. Use awk to pull a single field out of history or df -h.
4. Use find piped into xargs to run wc -l across every file in a folder at once.

Recap

• cut slices lines by character position (-c) or by delimiter and field (-d/-f).
• sed finds and replaces text; add -i to edit the file in place instead of just printing the result.
• awk prints specific fields from delimited or tabular text, using -F for the delimiter and $1, $2, etc. for fields.
• xargs takes a list of inputs and runs a command once per input, turning single-file commands into batch operations.

Next lesson: writing your own shell scripts, and editing files directly in the terminal with Vim.