shuf

Shuffle lines of a file

$ shuf [file]

$ shuf -n [5] [file]

$ shuf -i [1-100]

$ shuf -i [1-100] -n [10]

$ shuf -e [item1] [item2] [item3]

$ shuf -e -z [items] | xargs -0

$ shuf -r -n [10] -e [a] [b] [c]

$ shuf --random-source=[/dev/urandom] [file]

shuf [-n count] [-i range] [-e args] [-r] [options] [file]

shuf randomly permutes input lines. It reads input, shuffles the order, and outputs all lines in random sequence.
Without options, shuf outputs all input lines in random order. The -n option limits output to the first N lines after shuffling - effectively random sampling without replacement.
Input range mode (-i) generates sequential numbers and shuffles them. Combined with -n, this selects random numbers from a range. Useful for generating lottery numbers, random IDs, or sampling.
Echo mode (-e) shuffles command-line arguments instead of file lines. This enables shuffling small lists without creating temporary files.
Repeat mode (-r) enables sampling with replacement - the same line can appear multiple times in output. This is useful for bootstrap sampling or simulation.
The random source option ensures reproducible shuffling when given a deterministic source, useful for testing.

-n NUM, --head-count NUM

Output at most NUM lines.

Generate numbers from LO to HI.

Treat arguments as input lines.

Output can repeat (with replacement).

Use NUL as line delimiter.

Write to file instead of stdout.

Get random bytes from file.

Loads entire input into memory. Very large files may cause memory issues. Default randomness is good but not cryptographic. Without -n, outputs entire input. Different from sort -R which may group identical lines.

shuf is part of GNU coreutils, providing command-line random shuffling. While Unix systems had various random line selection tools, shuf provides comprehensive shuffling with range generation and sampling options.

sort(1), head(1), tail(1), sample(1)