How can I cut(1) camelcase words?

Question

Is there an easy way in Bash to split a camelcased word into its constituent words?

For example, I want to split aCertainCamelCasedWord into 'a Certain Camel Cased Word' and be able to select those fields that interest me. This is trivially done with cut(1) when the word separator is the underscore, but how can I do this when the word is camelcased?

score 27 · Accepted Answer · answered Mar 05 '09 at 14:18

27

sed 's/\([A-Z]\)/ \1/g'

Captures each capital letter and substitutes a leading space with the capture for the whole stream.

$ echo "aCertainCamelCasedWord" | sed 's/\([A-Z]\)/ \1/g'
a Certain Camel Cased Word

answered Mar 05 '09 at 14:18

Judge Maygarden

26,961
9
82
99

how do I output the words each line instead of space separated? – John Chen Feb 17 '20 at 12:35

score 3 · Answer 2 · edited Apr 02 '19 at 15:51

This solution works if you need to not split up words that are all caps. For example, using the top answer you'll get:

$ echo 'FAQPage' | sed 's/\([A-Z]\)/ \1/g' 
F A Q Page

But instead with my solution, you'll get:

$ echo 'FAQPage' | sed 's/\([A-Z][^A-Z]\)/ \1/g'
FAQ Page

Note: This does not work correctly when there is a second instance of multiple uppercase words, for example:

$ echo 'FAQPageOneReplacedByFAQPageTwo' | sed 's|\([A-Z][^A-Z]\)| \1|g'
FAQ Page One Replaced ByFAQ Page Two

score 2 · Answer 3 · edited May 23 '17 at 12:25

This answer does not work correctly when there is a second instance of multiple uppercase

echo 'FAQPageOneReplacedByFAQPageTwo' | sed 's|\([A-Z][^A-Z]\)| \1|g'
FAQ Page One Replaced ByFAQ Page Two

So and additional expression is required for that

 echo 'FAQPageOneReplacedByFAQPageTwo' | sed -e 's|\([A-Z][^A-Z]\)| \1|g' -e 's|\([a-z]\)\([A-Z]\)|\1 \2|g'
 FAQ Page One Replaced By FAQ Page Two

score 0 · Answer 4 · answered Mar 06 '09 at 09:55

Pure Bash:

name="aCertainCamelCasedWord"

declare -a word                                 # the word array

counter1=0                                      # count characters
counter2=0                                      # count words

while [ $counter1 -lt ${#name} ] ; do
  nextchar=${name:${counter1}:1}
  if [[ $nextchar =~ [[:upper:]] ]] ; then
    ((counter2++))
    word[${counter2}]=$nextchar
  else
    word[${counter2}]=${word[${counter2}]}$nextchar
  fi
  ((counter1++))
done

echo -e "'${word[@]}'"

Interesting, yet much more verbose. Use the right tool for the job I say! ;) — Judge Maygarden, Mar 19 '09 at 15:15

How can I cut(1) camelcase words?

4 Answers4

Linked