cheatsheet

gzip Command in Linux: Compress, Decompress, and Inspect .gz Files

gzip compresses single files with the DEFLATE algorithm, replacing each source with a .gz file by default. gunzip and zcat decompress; use -k to keep originals, -c to write to stdout, and -r for directory trees.

Published

Updated

Read time 9 min read

Reviewed byDeepak Prasad

gzip Command in Linux: Compress, Decompress, and Inspect .gz Files
About gzip compresses single files with the DEFLATE algorithm, replacing each source with a .gz file by default. gunzip and zcat decompress; use -k to keep originals, -c to write to stdout, and -r for directory trees.
Tested on Ubuntu 25.04 (Plucky Puffin); gzip 1.13; kernel 7.0.0-27-generic
Package gzip
Man page gzip(1)
Privilege user
Distros

gzip ships on virtually every Linux distro (GNU gzip).

Directory archives: combine with tar (tar czf archive.tar.gz dir/). Alternatives: bzip2, xz.
Related guide

gzip — quick reference

Compression

Compress one or more files. By default each source file is removed after a .gz is written in the same directory.

When to use Command
Compress a file (replaces original with .gz) gzip file.txt
Compress and keep the original file gzip -k file.txt
Write compressed data to stdout (leave source on disk) gzip -c file.txt > file.txt.gz
Overwrite an existing .gz without prompting gzip -f file.txt
Fastest compression (less ratio) gzip -1 file.txt
Best compression (slowest) gzip -9 file.txt
Compress every file under a directory tree gzip -r directory/
Rsync-friendly compression (extra sync-friendly flush points) gzip --rsyncable -c file.txt > file.txt.gz
Do not save original name/timestamp in the gzip header gzip -n file.txt
Save original name and timestamp (default) gzip -N file.txt
Use a custom suffix instead of .gz gzip -S .gz_backup file.txt

Decompression

gunzip and zcat are links to the same gzip binary — they select decompress mode by invocation name.

When to use Command
Decompress in place (removes .gz) gzip -d file.txt.gz
Same as above gunzip file.txt.gz
Decompress to stdout (keep .gz on disk) gzip -dc file.txt.gz
Same as above zcat file.txt.gz
Decompress from a pipe zcat file.txt.gz | head

Inspection and integrity

When to use Command
List compressed size, uncompressed size, and ratio gzip -l file.txt.gz
Test integrity without decompressing gzip -t file.txt.gz
Verbose compression statistics gzip -v file.txt
Suppress warnings gzip -q file.txt
Show brief usage gzip --help
Show package version gzip --version

gzip — command syntax

Synopsis from gzip --help on Ubuntu 25.04 (gzip 1.13):

text 
gzip [OPTION]... [FILE]...

With no FILE, or when FILE is -, gzip reads standard input. By default compression replaces each input file with a .gz in the same directory. Decompression modes are selected with -d or by calling gunzip / zcat.

gzip — command examples

Essential Compress a file and replace the original

The default action compresses file.txt and leaves only file.txt.gz beside it — the usual pattern for log rotation.

Run the commands:

bash 
echo 'compress me' > /tmp/glc-gzip-test/sample.txt
gzip -v /tmp/glc-gzip-test/sample.txt
ls -la /tmp/glc-gzip-test/sample.txt*

Sample output:

text 
/tmp/glc-gzip-test/sample.txt:	-16.7% -- replaced with /tmp/glc-gzip-test/sample.txt.gz
-rw-r--r-- 1 root root 43 Jul  1 19:06 /tmp/glc-gzip-test/sample.txt.gz

-v prints the compression ratio. The original plain file is gone unless you passed -k.

Essential Compress with -k and keep the source file

Use -k when you still need the uncompressed file on disk — common before uploading a .gz copy while keeping the live log.

Run the command:

bash 
echo 'compress me' > /tmp/glc-gzip-test/keepme.txt
gzip -k /tmp/glc-gzip-test/keepme.txt
ls -la /tmp/glc-gzip-test/keepme.txt*

Sample output:

text 
-rw-r--r-- 1 root root 12 Jul  1 19:06 /tmp/glc-gzip-test/keepme.txt
-rw-r--r-- 1 root root 43 Jul  1 19:06 /tmp/glc-gzip-test/keepme.txt.gz

Both files remain. Decompress later with gunzip or read without extracting using zcat.

Essential Compress or decompress through stdout

-c writes to standard output and never removes the source — ideal for pipes and when you want a .gz in a different directory.

Compress to a new path:

bash 
gzip -c /tmp/glc-gzip-test/keepme.txt > /tmp/glc-gzip-test/stdout-copy.gz

Decompress without touching the archive:

bash 
gzip -dc /tmp/glc-gzip-test/keepme.txt.gz
echo 'pipe test' | gzip | gunzip

Sample output:

text 
compress me
pipe test

zcat behaves like gzip -dc for reading .gz files in pipelines.

Common Inspect a .gz file without fully decompressing

Before shipping archives or after a partial download, -l shows sizes and -t verifies the gzip wrapper and checksum.

Run the commands:

bash 
gzip -l /tmp/glc-gzip-test/keepme.txt.gz
gzip -t /tmp/glc-gzip-test/keepme.txt.gz && echo 'integrity OK'

Sample output:

text 
compressed        uncompressed  ratio uncompressed_name
                 43                  12 -16.7% /tmp/glc-gzip-test/keepme.txt
integrity OK

A silent exit from gzip -t means the file is valid. Non-zero exit and a message mean the archive is truncated or corrupt.

Common Compress every file under a directory

-r walks a directory and gzips each regular file it finds — it does not create one tar-style archive; you get one .gz per file.

Run the commands:

bash 
mkdir -p /tmp/glc-gzip-test/sub
echo 'another' > /tmp/glc-gzip-test/sub/nested.txt
gzip -r /tmp/glc-gzip-test/sub
ls -la /tmp/glc-gzip-test/sub/
zcat /tmp/glc-gzip-test/sub/nested.txt.gz

Sample output:

text 
total 4
drwxr-xr-x 2 root root  60 Jul  1 19:06 .
drwxr-xr-x 3 root root 140 Jul  1 19:06 ..
-rw-r--r-- 1 root root  39 Jul  1 19:06 nested.txt.gz
another

For a single downloadable bundle, use tar (tar czf archive.tar.gz dir/) instead of gzip -r.

Common Decompress and remove the .gz file

gunzip restores the uncompressed file and deletes the .gz by default — the mirror of plain gzip.

Run the commands:

bash 
cp /tmp/glc-gzip-test/keepme.txt /tmp/glc-gzip-test/restore.txt
gzip /tmp/glc-gzip-test/restore.txt
ls /tmp/glc-gzip-test/restore.txt.gz
gunzip /tmp/glc-gzip-test/restore.txt.gz
ls /tmp/glc-gzip-test/restore.txt

Sample output:

text 
/tmp/glc-gzip-test/restore.txt.gz
/tmp/glc-gzip-test/restore.txt

Use gunzip -k or gzip -dk to keep the .gz after decompression.

Common Overwrite an existing output file

Without -f, gzip refuses to clobber an existing .gz or an uncompressed file that would be overwritten on decompress.

Run the command (plain file already exists):

bash 
gzip -f /tmp/glc-gzip-test/keepme.txt

Sample output:

text 
(no message — exit 0; keepme.txt.gz refreshed)

Verify:

bash 
ls -la /tmp/glc-gzip-test/keepme.txt*

If you see already exists; not overwritten, add -f or remove the conflicting file first.

Advanced Detect a corrupt or non-gzip file

gzip -t is the quick integrity check. Feeding a plain text file renamed to .gz fails immediately.

Run the command:

bash 
echo 'not gzip data' > /tmp/glc-gzip-test/corrupt.gz
gzip -t /tmp/glc-gzip-test/corrupt.gz

Sample output:

text 
gzip: /tmp/glc-gzip-test/corrupt.gz: not in gzip format

Exit status is non-zero. Re-transfer the file or restore from backup; partial downloads and disk errors often produce this error.

Advanced Trade speed for ratio with -1 and -9

Levels -1 (fast) through -9 (best) tune compression effort. Default is -6. For repetitive logs the size difference is often modest.

Run the commands:

bash 
echo 'repeat repeat repeat repeat' > /tmp/glc-gzip-test/level.txt
gzip -1 -c /tmp/glc-gzip-test/level.txt | wc -c
gzip -9 -c /tmp/glc-gzip-test/level.txt | wc -c

Sample output (byte counts vary slightly by content):

text 
45
45

On tiny files both levels can match. On large logs, -9 usually wins a few percent at the cost of CPU time — benchmark once on your data if batch compression runs for hours.

gzip — when to use / when not

Use gzip when Use something else when
  • You need fast, universal single-file compression with a .gz suffix
  • You are compressing logs, JSON exports, or text before upload or rotation
  • You want streaming compression in a pipe (gzip -c)
  • You are building .tar.gz archives with tar czf
  • You need a quick integrity check on an existing .gz (gzip -t)
  • You need better compression ratio on large files and can spend more CPU → xz or bzip2
  • You need one archive containing many files and paths → tar (optionally with gzip via z)
  • You need zip compatibility with Windows tools → zip / zip folder guide
  • You need random access inside a compressed bundle → zip or an uncompressed tar
  • You need parallel compression of huge trees → pigz, lbzip2 (extra packages)

gzip vs bzip2 and xz

gzip bzip2 xz
Speed Fastest of the three Slower Slowest (best ratio)
Ratio Good on text Better Best
Typical extension .gz .bz2 .xz
tar idiom tar czf tar cjf tar cJf

gzip remains the default for logs and tar.gz because almost every Linux system ships it and decompression is cheap.

Nearby tools for archiving, stronger compression, and bundled directories.

Command One line
gzip Single-file .gz compression (this page)
gunzip / zcat Decompress .gz to disk or stdout (same binary)
tar Bundle many files; add z for gzip
bzip2 Higher ratio, .bz2 suffix
xz Strongest common ratio, .xz suffix

Browse the full index on the Linux commands cheat sheet.

gzip — interview corner

What does the gzip command do in Linux?

gzip compresses one file at a time using the DEFLATE algorithm. By default it deletes the original and leaves filename.gz in the same directory. It does not archive directory structure by itself — pair it with tar for archive.tar.gz.

Common companions on the same binary: gunzip (decompress in place), zcat (decompress to stdout).

A strong answer is:

"gzip compresses individual files to .gz, replacing the source by default. For folders I use tar czf; gunzip and zcat decompress."

How do you compress a file but keep the original?

Pass -k (keep):

bash 
gzip -k file.txt

Or write to stdout and redirect so the source is never removed:

bash 
gzip -c file.txt > file.txt.gz

-c is the usual pattern in pipelines; -k is simpler when both files should live in the same directory.

A strong answer is:

"I use gzip -k to keep the plain file, or gzip -c with redirection when the .gz should land elsewhere or feed a pipe."

What is the difference between zcat and gunzip?

Both invoke the gzip binary in decompress mode.

  • gunzip writes the uncompressed file to disk and removes the .gz (unless -k).
  • zcat (like gzip -dc) writes to stdout and leaves the .gz untouched — like cat for compressed files.
bash 
zcat file.txt.gz | less
gunzip file.txt.gz    # creates file.txt, removes file.txt.gz

A strong answer is:

"gunzip materializes the plain file on disk; zcat streams decompressed bytes to stdout for pipes and viewers."

How do you verify a .gz file is not corrupted?

Run:

bash 
gzip -t file.txt.gz

Exit code 0 means the gzip container and CRC check passed. No output is normal. Use gzip -l when you also want compressed and uncompressed sizes.

A strong answer is:

"gzip -t validates integrity without unpacking; I use it after transfers and before deleting the only uncompressed copy."

Does gzip -r create one archive of a directory?

No. gzip -r compresses each file under a directory separately — you end up with many .gz files, not one bundle.

For one archive:

bash 
tar czf archive.tar.gz directory/

A strong answer is:

"gzip -r gzips every file in a tree individually; for a single archive I use tar czf or zip."

Troubleshooting

Symptom Likely cause Fix
not in gzip format File is not gzip, truncated, or wrong extension Re-download; verify with file foo.gz; test with gzip -t
already exists; not overwritten Target file or .gz present gzip -f or remove the conflicting file
Original vanished after compress Default replace behaviour Use gzip -k or gzip -c next time
unexpected end of file Truncated download or partial write Re-transfer; check disk space during write
gzip: stdin: not in gzip format Piped input is not gzip Ensure upstream outputs gzip; use zcat only on .gz data
Huge directory with gzip -r Many separate .gz files, hard to ship Use tar czf for one archive

References

Further reading — man page and related compression guides.

Rohan Timalsina

is a technical writer and Linux enthusiast who writes practical guides on Linux commands and system administration. He focuses on simplifying complex topics through clear explanations.