CODEWITHSUNDEEP
bashshellsed
rodee
rodee

Reputation: 3151

How to convert a line into camel case?

This picks all the text on single line after a pattern match, and converts it to camel case using non-alphanumeric as separator, remove the spaces at the beginning and at the end of the resulting string, (1) this don't replace if it has 2 consecutive non-alphanumeric chars, e.g "2, " in the below example, (2) is there a way to do everything using sed command instead of using grep, cut, sed and tr?

$ echo " hello
world
  title:  this is-the_test string with number 2, to-test CAMEL String
end! " | grep -o 'title:.*' | cut -f2 -d: | sed -r 's/([^[:alnum:]])([0-9a-zA-Z])/\U\2/g' | tr -d ' '
ThisIsTheTestStringWithNumber2,ToTestCAMELString

Upvotes: 1

Views: 683

Answers (1)

Wiktor Stribiżew
Wiktor Stribiżew

Reputation: 626691

To answer your first question, change [^[:alnum:]] to [^[:alnum:]]+ to mach one ore more non-alnum chars.

You may combine all the commands into a GNU sed solution like

sed -En '/.*title: *(.*[[:alnum:]]).*/{s//\1/;s/([^[:alnum:]]+|^)([0-9a-zA-Z])/\U\2/gp}'

See the online demo

Details

  • -En - POSIX ERE syntax is on (E) and default line output supressed with n
  • /.*title: *(.*[[:alnum:]]).*/ - matches a line having title: capturing all after it up to the last alnum char into Group 1 and matching the rest of the line
  • {s//\1/;s/([^[:alnum:]]+|^)([0-9a-zA-Z])/\U\2/gp} - if the line is matched,
    • s//\1/ - remove all but Group 1 pattern (received above)
    • s/([^[:alnum:]]+|^)([0-9a-zA-Z])/\U\2/ - match and capture start of string or 1+ non-alnum chars into Group 1 (with ([^[:alnum:]]+|^)) and then capture an alnum char into Group 2 (with ([0-9a-zA-Z])) and replace with uppercased Group 2 contents (with \U\2).

Upvotes: 1

Related Questions