gzip

What it does

gzip compresses a single file using the DEFLATE algorithm, producing a .gz file and, by default, deleting the original. It's the classic Unix compression tool, distinct from archivers like tar in that it only compresses one file at a time and has no concept of bundling multiple files or directories — that's why tar and gzip are so often used together (tar -czf). Its companion tools gunzip, zcat, and zmore handle decompression and inspection of .gz files.

Beginner examples

• gzip file.txt — compress file.txt into file.txt.gz, deleting the original
• gzip -k file.txt — compress but keep the original file too
• gzip -d file.txt.gz — decompress back to file.txt
• gzip -9 file.txt — compress using maximum compression level

gzip -k access.log

Advanced examples

• Compress every file in a directory tree recursively: gzip -r logs/
• Combine with tar to compress an entire directory into one archive: tar -czf site_backup.tar.gz site/
• Compress to stdout so you can pipe the result elsewhere without touching the source file: gzip -c database_dump.sql > database_dump.sql.gz
• Check the compression ratio and sizes of an existing .gz file without decompressing it: gzip -l archive.tar.gz
• Fine-tune the speed/size tradeoff for very large files where -9 is too slow: gzip -1 huge_dataset.csv

tar -cf - large_dir/ | gzip -9 > large_dir_backup.tar.gz

Common mistakes

• Expecting gzip to compress a whole directory into one file the way zip does — gzip only compresses single files; use tar -czf for directories.
• Forgetting -k and losing the uncompressed original when you actually wanted to keep both versions.
• Re-compressing an already-compressed format (like a .jpg, .mp4, or .zip) and being surprised the .gz file is barely smaller — DEFLATE gets little benefit from data that's already compressed.
• Confusing gzip -d file.gz with simply running gunzip file.gz — they're equivalent, but mixing up the flag can lead to confusion about which mode you're in.

Tips

• Use zcat file.gz to view a compressed file's contents without permanently decompressing it to disk.
• gzip -l shows the compression ratio, which is a quick sanity check on whether compressing a particular file type is even worthwhile.
• For maximum compatibility with other tools expecting .tar.gz, always compress the tar archive as a whole (tar -czf) rather than gzip-ing many individual files inside a directory.

Best practices

• Use -k by default when compressing anything you might need to reference again quickly, since decompression, while fast, is an extra step you can avoid.
• Prefer tar -czf over manually gzip-ing many files when the goal is to package and compress a whole directory — it preserves structure and permissions in one archive.
• Reserve -9 for cases where storage or bandwidth is the priority and CPU time is cheap; for frequently-compressed data (like rotated logs), the default level is usually a better time/space tradeoff.

Real-world use cases

• Compressing rotated log files to save disk space: gzip /var/log/app.log.1.
• Shrinking a database dump before transferring it off-server: gzip -9 database_backup.sql.
• Building a compressed tarball of a project for distribution: tar -czf project-v1.0.tar.gz project/.

Common interview questions

• Does gzip delete the original file? Yes, by default; use -k to keep it, or -c to write compressed output to stdout without touching the source at all.
• Why can't gzip compress an entire directory into one archive? gzip only operates on individual files and has no archive format of its own; combining multiple files requires tar first, typically piped or invoked together as tar -czf.
• What's the difference between gzip -d file.gz and gunzip file.gz? They're functionally identical — gunzip is essentially gzip -d, both decompress the file and remove the .gz extension.