6

I have text that goes like this:

I am happy. I am here. How are you, Meg? 

I want this to be:

I am happy.
I am here.
How are you, Meg? 

For full stops I tried

tr -s '. ' '\n' <file.txt >out.txt

But it's not working.

7

As far as I know tr only works with single characters and ". " is a string not a character, so it is possible to do what you want by using sed or awk, for example:

sed -e "s/\. /\n/g" file.txt > out.txt
  • He said use tr. – argle Jan 3 '17 at 12:11
  • 9
    @argle Well, he can't. – Kusalananda Jan 3 '17 at 12:16
  • 2
    I know. I just didn't realize xy answers were allowed even against explicit constraints. This could have been my answer. So I was protesting. :P – argle Jan 3 '17 at 12:25
14

You don't. tr is not designed for that. It's designed to transliterate a set of single characters into another set of single characters, e.g., A-Z into a-z. Using tr with .␣ (a dot and a space) and \n will replace all dots and spaces with newlines.

Use (GNU) sed instead:

$ echo 'I am happy. I am here. How are you, Meg?' | sed 's/\([!.?]\) /\1\n/g'
I am happy.      
I am here.       
How are you, Meg?

The sed editing script here will replace all occurrences of !, . or ? that are followed by a space, by the same character and a newline.

  • 1
    I'd like to emphasize single characters since OP seems to believe tr will replace TWO characters '. ' with just one character '\n' - Instead tr replaces both '.' and ' ' with '\n'. – ihato Jan 3 '17 at 12:46
  • @Yanevski Emphasis added. – Kusalananda Jan 3 '17 at 12:57
0

As has been mentioned , you can't achieve what you want with tr due to the nature of tr. But there are many other alternatives. For instance, Python 2.

As one-liner we could do this:

python -c "import sys;print '\n'.join([ j.strip() for l in sys.stdin.readlines() for j in  l.rstrip().split('.')])" < input.txt

As a script , it would be like so:

#!/usr/bin/env python
import sys

sentences = []
for line in sys.stdin:
    for sentence in line.rstrip().split('.'):
        sentences.append(sentence.strip())

print "\n".join(sentences)

And used as so:

$ ./split_to_lines.py  < input.txt                                                                    
I am happy
I am here
How are you, Meg?

Nothing particularly complex is going on here - we're merely splitting all lines in file into sentences at . and then strip leading and trailing space for each sentence. All that goes into a list, which is then reassembled via .join() function.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.