sort Command in Linux: Sort Lines of Text

sort is a command-line utility that sorts lines of text from files or standard input and writes the result to standard output. By default, sort arranges lines alphabetically, but it supports numeric sorting, reverse order, sorting by specific fields, and more.

This guide explains how to use the sort command with practical examples.

sort Command Syntax #

The syntax for the sort command is as follows:

sort [OPTIONS] [FILE]...

If no file is specified, sort reads from standard input. When multiple files are given, their contents are merged and sorted together.

Sorting Lines Alphabetically #

By default, sort sorts lines in ascending alphabetical order. To sort a file, pass the file name as an argument:

sort file.txt

For example, given a file with the following content:

banana
apple
cherry
date

Running sort produces:

apple
banana
cherry
date

The original file is not modified. sort writes the sorted output to standard output. To save the result to a file, use output redirection:

sort file.txt > sorted.txt

Sorting in Reverse Order #

To sort in descending order, use the -r (--reverse) option:

sort -r file.txt

date
cherry
banana
apple

Sorting Numerically #

Alphabetical sorting does not work correctly for numbers. For example, 10 would sort before 2 because 1 comes before 2 alphabetically. Use the -n (--numeric-sort) option to sort numerically:

sort -n numbers.txt

Given the file:

10
2
30
5

The output is:

2
5
10
30

To sort numerically in reverse (largest first), combine -n and -r:

sort -nr numbers.txt

30
10
5
2

Sorting by a Specific Column #

By default, sort uses the entire line as the sort key. To sort by a specific field (column), use the -k (--key) option followed by the field number.

Fields are separated by whitespace by default. For example, to sort the following file by the second column (file size):

report.pdf 2500
notes.txt 80
archive.tar 15000
readme.md 200

sort -k2 -n files.txt

notes.txt 80
readme.md 200
report.pdf 2500
archive.tar 15000

To use a different field separator, specify it with the -t option. For example, to sort /etc/passwd by the third field (UID), using : as the separator:

sort -t: -k3 -n /etc/passwd

Removing Duplicate Lines #

To output only unique lines (removing adjacent duplicates after sorting), use the -u (--unique) option:

sort -u file.txt

Given a file with repeated entries:

apple
banana
apple
cherry
banana

apple
banana
cherry

This is equivalent to running sort file.txt | uniq.

Ignoring Case #

By default, sort is case-sensitive. Uppercase letters sort before lowercase. To sort case-insensitively, use the -f (--ignore-case) option:

sort -f file.txt

Checking if a File Is Already Sorted #

To verify that a file is already sorted, use the -c (--check) option. It exits silently if the file is sorted, or prints an error and exits with a non-zero status if it is not:

sort -c file.txt

sort: file.txt:2: disorder: banana

This is useful in scripts to validate input before processing.

Sorting Human-Readable Sizes #

When working with output from commands like du, sizes are expressed in human-readable form such as 1K, 2M, or 3G. Standard numeric sort does not handle these correctly. Use the -h (--human-numeric-sort) option:

du -sh /var/* | sort -h

4.0K /var/backups
16K /var/cache
1.2M /var/log
3.4G /var/lib

Combining sort with Other Commands #

sort is commonly used in pipelines with other commands.

To count and rank the most frequent words in a file, combine grep
, sort, and uniq:

grep -Eo '[[:alnum:]_]+' file.txt | sort | uniq -c | sort -rn

To display the ten largest files in a directory, combine sort with head
:

du -sh /var/* | sort -rh | head -10

To extract and sort a specific field from a CSV using cut
:

cut -d',' -f2 data.csv | sort

Quick Reference #

Troubleshooting #

Numbers sort in wrong order
You are using alphabetical sort on numeric data. Add the -n option to sort numerically: sort -n file.txt.

Human-readable sizes sort incorrectly
Sizes like 1K, 2M, and 3G require -h (--human-numeric-sort). Standard -n only handles plain integers.

Uppercase entries appear before lowercase
sort is case-sensitive by default. Use -f to sort case-insensitively, or use LC_ALL=C sort to enforce byte-order sorting.

Output looks correct on screen but differs from file
Output redirection with > truncates the file before reading. Never redirect to the same file you are sorting: sort file.txt > file.txt will empty the file. Use a temporary file or the sponge utility from moreutils.

FAQ #

Does sort modify the original file?
No. sort writes to standard output and never modifies the input file. Use sort file.txt > sorted.txt or sort -o file.txt file.txt to save the output.

How do I sort a file and save it in place?
Use the -o option: sort -o file.txt file.txt. Unlike > file.txt, the -o option is safe to use with the same input and output file.

How do I sort by the last column?
Use a negative field number with the -k option: sort -k-1 sorts by the last field. Alternatively, use awk to reorder columns before sorting.

What is the difference between sort -u and sort | uniq?
sort -u is slightly more efficient as it removes duplicates during sorting. sort | uniq is more flexible because uniq supports options like -c (count occurrences) and -d (show only duplicates).

How do I sort lines randomly?
Use sort -R (random sort) or the shuf command: shuf file.txt. Both produce a random permutation of the input lines.

Conclusion #

The sort command is a versatile tool for ordering lines of text in files or command output. It is most powerful when combined with other commands like grep
, cut
, and head
in pipelines.

If you have any questions, feel free to leave a comment below.