NAME

     unsort — reorder lines in a file in semirandom ways

SYNOPSIS

     unsort [-hvrpncmMsz0l] [--help] [--version] [--random] [--heuristic] [--identity]
            [--filenames[=profile]] [--separator sep] [--concatenate] [--merge] [--merge-random]
            [--seed integer] [--zero-terminated] [--null] [--linefeed] [file ...]

DESCRIPTION

     unsort prints the lines in the input files (or standard input) in semi-random order.
     Available algorithms are a Mersenne Twister based PRNG and a heuristic algorithm that aims
     to create a subjective even distribution.

   Command line options
     -h, --help
             Display a concise summary of the available options and argument syntax.

     -v, --version
             Display version and copyright information.

     -r, --random
             Use the Mersenne Twister based randomization algorithm.

     -p, --heuristic
             Use the heuristic “shuffling” algorithm which permutes the lines in such a way that
             they're spread more or less evenly in the output.  This is the default.

     -n, --identity
             Do not reorder lines in the input. Useful if you just want to merge the files.

     -f, --filenames[=profile]
             Assumes the input has a structure similar to filenames and apply randomization
             seperately to each directory level.

             An example would be a list of mp3 files where the top level is the genre, the second
             level is the name of the artist and the third level is the name of the album.  The
             --filenames option allows you to randomize the list with different randomization
             strategies for each of the artist, album and track levels.

             Two types of randomization are supported: simple and proportional.  Simple
             randomization picks one entry from each directory at that level in a round-robin
             fashion.  If one directory contains fewer entries than the others, it will run out
             and stop appearing at some point before the end of the output.

             Proportional randomization will spread the directory entries evenly over the output.

             The profile argument can be used to specify which randomization types to use.  It is
             a concatenation of the randomization types to use for each level.  Use 1 for simple
             randomization and n for proportional randomization.  If the input has more levels
             than specified, the last character determines the randomization type for all
             subsequent levels.

             For example, --filenames=n1n means: use proportional randomization for the genre,
             use simple randomization for the artists and use proportional randomization again
             for the album names and all levels below that.

     -S, --separator sep
             The directory separator to use with -f.  Defaults to ‘/’.

     -r, --concatenate
             Concatenate all input files then apply the shuffling algorithm to the result as a
             whole.

     -m, --merge
             Shuffle all input files seperately then merge the result.  Equal-sized files will be
             merged in the order in which they appear on the command line.

     -M, --merge-random
             Shuffle all input files seperately then merge the result.  Equal-sized files will be
             merged in random order.  This is the default.

     -s, --seed integer
             Use this integer as a seed, instead of random data from the environment.

     -z, --zero-terminated, -0, --null
             Lines are terminated with a \0 character.

     -l, --linefeed
             Lines are terminated with a \n character.  This is the default.

SEE ALSO

     sort(1)