[SOLVED] BASH - Formating - How to Cap all words but not the 's, 't etc

BW-userx · 12-12-2015, 02:21 PM

I have found two ways to cap all words in a string, but it causes it to leave me with words like this,

Code:

 Don'T

, and any other words that have the ' apostrophe in it the letter after it is also capitlized, which is not needed.

Code:

 echo "hi don't think's so" | sed -e  's/\<\(.\)\([A-Za-z]*\)\>/\u\1\L\2/g'
Hi Don'T Think'S So

and

Code:

file="$(echo -e "${file}" | tr "[A-Z]" "[a-z]" | sed -e "s/\b\(.\)/\u\1/g")"

which both do the same thing, leaving the letter just after that apostrophe also upper case which it is not suppose to be. every example loves to show how to upper case everything but not how to handle this situation. gezz -

How do I prevent that from happening?

HMW · 12-12-2015, 02:48 PM

Hi!

Real quick:

Code:

echo "hi don't think's so" | sed 's/\( [a-z]\)/\U\1/g;s/\(^.\)/\U\1/' 
Hi Don't Think's So

There's room for improvement I'm sure, but I'm tired...

BW-userx · 12-12-2015, 03:07 PM

Quote:

Originally Posted by HMW

Hi!

Real quick:

Code:

echo "hi don't think's so" | sed 's/\( [a-z]\)/\U\1/g;s/\(^.\)/\U\1/' 
Hi Don't Think's So

There's room for improvement I'm sure, but I'm tired...

I know that feeling --

take a nap -- cheers

grail · 12-12-2015, 09:31 PM

I am guessing we need the \L because we might have mixed cases elsewhere?

Code:

echo "hi don't think's so" | sed -r 's/([\b ]?)(.)([a-zA-Z]*)/\1\U\2\L\3/g'

Aia · 12-12-2015, 10:58 PM

Code:

echo "hi don't think's so \"hello\" yoOOUUuu" | perl -pe 's/([\w\x27]+)/\u\L$1/g'

Code:

Hi Don't Think's So "Hello" Yooouuuu

joel2001k · 12-13-2015, 02:42 AM

http://tldp.org/LDP/Bash-Beginners-G...ect_08_02.html

BW-userx · 12-13-2015, 02:29 PM

Quote:

Originally Posted by grail

I am guessing we need the \L because we might have mixed cases elsewhere?

Code:

echo "hi don't think's so" | sed -r 's/([\b ]?)(.)([a-zA-Z]*)/\1\U\2\L\3/g'

I was using the other example then it just got to me , I couldn't take it no longer

, seeing that letter just after the apostrophe being capped everytime, so I looked up trying to find a way to fix it, the came across that other example. I too got it off the interNet I decided to give it a try becuase that L insures uniformity within the strings, the first char is CAP and the 2 and after is lower case, then it hits that apostrophe then throws everything off, capping the letter just after it. Then my brain got to hurtting trying to figure out how to mask that apostrophe so that it'd see it as an apostrophe and not that other way it sees it- I'm Jargon handycapped

BW-userx · 12-13-2015, 02:34 PM

Quote:

Originally Posted by Aia

Code:

echo "hi don't think's so \"hello\" yoOOUUuu" | perl -pe 's/([\w\x27]+)/\u\L$1/g'

Code:

Hi Don't Think's So "Hello" Yooouuuu

so what is that code saying in english? where is it telling the BAsh to not see that apostrophe as whatever it sees it as other then part of a sentance? it needs to be told to see the apostrophe as part of the sentense. what part of that code says that, so I'll better unsderstand it. 2. I see you're using escape \ within the text to read, where as what I am doing is dynamic reading file names off a loop (didn't add that info, just showed a command line example, my bag) that'd require more code to have to put excapes within the file name before it is processed, yes?

Aia · 12-13-2015, 04:26 PM

Quote:

Originally Posted by BW-userx

so what is that code saying in english?

Code:

perl -pe 's/([\w\x27]+)/\u\L$1/g'

\w : any alphanumeric character
\x27 : single quote in hex, it could be \047 in octal, since the command line would not allow to just place ' without a lot of pain
[\w\x27] : any of these two
[\w\x27]+ : any of these two at least once, but possible more. This will match any word with or without apostrophe.
([\w\x27]+) : capture this group and save it as $1

\u\L$1
$1 : the captured string i.e. word
\L : lower any char until the \E is given, none was given
\u : capitalize the first character

/g : repeat the whole regex as many times as possible.

danielbmartin · 12-13-2015, 05:27 PM

Consider using this short test file which is more challenging.

Code:

hi don't think's so
  you've got to be joking!
My country 'tis of thee
'tis the wind and nothing more!

Daniel B. Martin

BW-userx · 12-13-2015, 05:35 PM

Quote:

Originally Posted by danielbmartin

Consider using this short test file which is more challenging.

Code:

hi don't think's so
  you've got to be joking!
My country 'tis of thee
'tis the wind and nothing more!

Daniel B. Martin

my copied Internet code

Code:

[userx@voided ~]$ echo "hi don't think's so
>   you've got to be joking!
> My country 'tis of thee
> 'tis the wind and nothing more!" | sed -e  's/\<\(.\)\([A-Za-z]*\)\>/\u\1\L\2/g'
Hi Don'T Think'S So
  You'Ve Got To Be Joking!
My Country 'Tis Of Thee
'Tis The Wind And Nothing More!

@grail code

Code:

[userx@voided ~]$ echo "hi don't think's so
>   you've got to be joking!
> My country 'tis of thee
> 'tis the wind and nothing more!" |  sed -r 's/([\b ]?)(.)([a-zA-Z]*)/\1\U\2\L\3/
g'

Hi Don't Think's So
  you've Got To Be Joking!
My Country 'tis Of Thee
'tis The Wind And Nothing More!

what are we looking 4 here? apples or peaches and cheeries?

Aia · 12-13-2015, 05:57 PM

If you feel 'tis must be 'Tis, please use the following:

Code:

perl -pe 's/(\w+\x27?(\w+)?)/\u\L$1/g' test.file

Code:

Hi Don't Think's So
  You've Got To Be Joking!
  My Country 'Tis Of Thee
  'Tis The Wind And Nothing More!

  This Line Below Is To Test Single Letters
  A B C'

danielbmartin · 12-13-2015, 07:14 PM

Quote:

Originally Posted by Aia

If you feel 'tis must be 'Tis, please use the following:

Code:

perl -pe 's/(\w+\x27?(\w+)?)/\u\L$1/g' test.file

Code:

Hi Don't Think's So
  You've Got To Be Joking!
  My Country 'Tis Of Thee
  'Tis The Wind And Nothing More!

  This Line Below Is To Test Single Letters
  A B C'

Nitpick: this result has leading blanks where the InFile did not.

Daniel B. Martin

Aia · 12-13-2015, 07:55 PM

Quote:

Originally Posted by danielbmartin

Nitpick: this result has leading blanks where the InFile did not.

Daniel B. Martin

I am afraid you are not nitpick in this instance, but rather leading to irrelevant. Do not get offended.