Get all possible combinations of a word in lower/uppercase letters

Question

I want to write a bash script to print all possible lower and upper case permutations of a certain word, e.g. harley:

harley
harleY
harlEy
harLey
...
HARLey
HARLEy
HARLEY

My naive solution is to write a n-th (n is len(word)) nested for-loop for this specific word:

#!/bin/bash
for a in {h,H}; do
    for b in {a,A}; do
    ...
    done
done

However, I would have to code the script for a different word again.

Is there a better way to accomplish this?

Answer 1

A slightly better solution:

echo {h,H}{a,A}{r,R}{l,L}{e,E}{y,Y}

For full scalability:

echo harley \
| perl -nle 'print "echo ",
                    join "",map { "{" . lc . "," .uc ."}" } split //' \
| xargs -I {} bash -c "{}"

If you absolutely must have one word per line, go with

for w in {h,H}{a,A}{r,R}{l,L}{e,E}{y,Y};do echo $w;done

thanks to mattdm's comment

The corresponding scalable version would be:

echo harley \
| perl -nle 'print join "",map { "{" . lc . "," .uc ."}" } split //' \
| xargs -I {} bash -c 'for w in {};do echo $w;done'

For fun, try replacing "harley" with "supercalifragilisticexpialidocious" It's been 5 minutes and my computer is still crunching on this one and will probably never finish :)

Answer 2

eval echo $(echo "word" | sed 's/./{\U&,\L&}/g')

• sed 's/./{&,&}/g' would turn Foo into {F,F}{o,o}{o,o}, which would be pretty useless.  But add \U and \L and you get the upper and lower case of each letter; i.e., {F,f}{O,o}{O,o}.
• Then it’s a simple matter of using eval to tell the shell to expand the {X,x} brace sequences.

Answer 3

EDIT 2: This answer is wrong. It doesn't produce 2^n combinations as it should.

EDIT: I don't know why, but this solution is realy fast compared to the perl solution by @Joeseph R. It runs "Supercalifragilisticexpialidocious" in less than 0.3 seconds!

Here's my crack at it:

#!/bin/bash

str=${1^^}  # convert to uppercase
len=${#str} # get length of string

for ((perm=0; perm <= len; perm++)); do
    for ((i=0; i <= len; i++)); do
        lower=${str,,}   # convert to lowercase

        # Uppercase n-th letter for permutation
        if [ $perm -gt 0 ]; then
            nth=${lower:perm-1}
            lower=$(echo ${lower:0:perm-1}${nth^})
        fi

        echo -n ${str:0:i} # print orig string from 0 to $i
        echo ${lower:i}    # print new string from $i to end
    done
done | sort -u

Running it:

$ ./permutations.sh hi
hi
hI
Hi
HI

$ ./permutations.sh harley
harley
harleY
harlEy
harLey
haRley
hArley
Harley
HarleY
HarlEy
HarLey
HaRley
HArley
HArleY
HArlEy
HArLey
HARley
HARleY
HARlEy
HARLey
HARLeY
HARLEy
HARLEY

Feel free to fork and modify it, I'm sure it can be optimized. https://gist.github.com/ryanmjacobs/4c02ad80f833dee0c307

Answer 4

With zsh:

set -o extendedglob # usually in ~/.zshrc

word=Harley
eval print -rl ${word//(#m)?/{$MATCH:l,$MATCH:u}}

Answer 5

I wrote a few simple functions based on the best answer that will do output in the desired format.

to_lower () 
{ 
    tr '[[:upper:]]' '[[:lower:]]' <<< $@
}
to_upper () 
{ 
    tr '[[:lower:]]' '[[:upper:]]' <<< $@
}
generate_permutations () 
{ 
    local perm="$@";
    perl -e '@foo = split(//, $ARGV[0]);foreach my $c (@foo){print "$c $c\n";}' ${perm} | while read l u; do
        echo "{$(to_lower ${l}),$(to_upper $u)}";
    done | tr -d '\n';
    echo
}

Usage/example:

$ generate_permutations foobar
{f,F}{o,O}{o,O}{b,B}{a,A}{r,R}

Answer 6

Improving on Joseph R's answer, this one shouldn't cause duplicates for characters that aren't a-z (improving processing time):

echo s.t.r | perl -nle 'print join "", map { /[a-zA-Z]/ ? "{" . lc . "," .uc ."}"  : $_ } split //' | xargs -I {} bash -c 'for w in {};do echo $w;done'
s.t.r
s.t.R
s.T.r
s.T.R
S.t.r
S.t.R
S.T.r
S.T.R