How can I count lines, words, or characters in a text file from the command line?
Quick answer
With GNU wc, use wc -l file.txt for newline characters, wc -w file.txt for whitespace-separated words, wc -m file.txt for characters, or wc -c file.txt for bytes. Combine options for several counts at once. Note that -l may not count a final line without a newline, and -m depends on the current locale. [S1]
The question, the details, the next step.
I have a text file and want a quick command-line count. Which commands should I use for lines, words, and characters?
For GNU wc, replace sample.txt below with your filename:
wc -l sample.txtcounts newline characters.wc -w sample.txtcounts words, defined as nonempty sequences separated by whitespace.wc -m sample.txtcounts characters according to the current locale.wc -c sample.txtcounts bytes.
To request all four counts at once, run wc -l -w -m -c sample.txt. GNU wc prints the counts followed by the filename. These details describe GNU wc; other systems or command versions may differ. [S1]
Could you show me a small example and how to read the result?
Assuming a shell where printf is available, create a small example file with two newline-terminated lines:
printf 'red fox\nblue\n' > sample.txt
Then run:
wc -l -w -m -c sample.txt
For GNU wc, the counts should correspond to 2 newline characters, 3 whitespace-separated words, 13 characters, and 13 bytes. The text is ASCII, so its character and byte counts match in this example. The displayed output also includes sample.txt; spacing can vary. [S1]
Why might wc -l give one fewer line than I see in a text editor? And is -m always the character count I mean?
wc -l counts newline characters, not every line-shaped piece of text. If a file has a final line without a newline, GNU wc -l does not count that trailing partial line, so the result can be lower than the number of lines you informally see. [S1]
-m counts characters according to the current locale, and GNU wc does not count encoding errors. It is not interchangeable with the byte count from -c. [S1] If you need a particular application’s definition of “character,” check that definition before comparing results. The rules described here are for GNU wc, not every platform or application.
What’s a practical way to choose the right count when I’m checking a file?
Start by deciding what you need: newline characters, whitespace-separated words, locale-based characters, or bytes. For an overview, use wc -l -w -m -c sample.txt; for one number, use the matching option alone. If the line total seems unexpected, check whether the final line ends with a newline. If character and byte totals differ, verify the locale and encoding relevant to your task rather than assuming one count is universally correct. [S1]
A reusable template is:
wc -l -w -m -c [path-to-file]
Replace the bracketed placeholder with the file’s path. This command is for GNU wc; check your platform’s documentation if its behavior or options matter.
A practical checklist
Choose whether you need newline, word, character, or byte counts.
Run the matching
wcoption, or combine options for an overview.Use
-cwhen you need bytes rather than a locale-based character count.If the line total looks low, check whether the final line has a newline.
Confirm your platform’s
wcbehavior if results must match another tool.
A few more useful answers
Does `wc -l` count a final line that has no newline?
GNU wc -l counts newline characters, so it does not count a trailing partial line without a newline as another newline. [S1]
What is the difference between `wc -m` and `wc -c`?
-m reports characters according to the current locale; -c reports bytes. GNU wc does not count encoding errors in its character count. [S1]
Does `wc -w` use the same word definition as every word processor?
Not necessarily. GNU wc treats a word as a nonempty sequence separated by whitespace, and whitespace interpretation is affected by locale. [S1]
Sources
References used for this guide.
- [S1] GNU Coreutils 9.11: wc — Print newline, word, and byte counts gnu.org · Checked