compress

Compress specific files

$ compress [path/to/file1 path/to/file2 ...]

$ compress -f [path/to/file1 path/to/file2 ...]

$ compress -b [bits]

$ compress -c [path/to/file]

$ compress -d [path/to/file]

$ compress -v [path/to/file]

compress [-dfvc] [-b maxbits] [file …]

-b maxbits
    Maximum bits per code (9-16, default 12)

-c
    Write compressed data to stdout (no .Z files)

-d
    Decompress instead of compressing

-f
    Force compression, even if file grows or is symlink

-v
    Verbose: print compression factor for each file

The compress command is an early Unix utility that reduces file sizes using the adaptive Lempel-Ziv-Welch (LZW) algorithm, a dictionary-based method developed in the early 1980s. It processes one or more input files, replacing each with a compressed version suffixed by .Z (e.g., file.txt becomes file.txt.Z). Original files are removed unless the -c option directs output to stdout.

Unlike modern tools like gzip, compress does not support tar archives or multi-file concatenation natively and produces lower compression ratios. It excels on text files, achieving 2-3x reduction typically. Decompression uses compress -d or the separate uncompress command; gzip -d also handles .Z files for compatibility.

Primarily historical, it's rarely used today due to patent issues with LZW (expired 2003) and superior alternatives. Available via ncompress package on Linux. Verbose mode reports compression stats, like percentage saved.

Obsolete; poor ratios vs gzip. Not default-installed. LZW patents delayed adoption. Avoid for new use.

compress -v file.txt
Replaces file.txt with file.txt.Z, shows ratio.

compress -dc file.Z
Outputs decompressed data to stdout.

Originated in 2.0BSD (1983), based on 1982 LZW paper by Welch et al. Ported to SVR3. Supplanted by gzip (1992) amid patent disputes.

gzip(1), uncompress(1), pack(1), bzip2(1), xz(1)