inserting spaces inbetween only capital letters that are directly adjacent to each other

user_16809190 0
Viewed 66

My goal is to add spaces in between letters of various abbreviations.

There are three conditions for abbreviations:

  1. abbreviations contain at least two or more letters.
  2. abbreviations are always capitalized.
  3. special character "/" is considered a capitalized letter

think of DNS, IP, TCP/IP, etc.

I would like to do text processing on them to turn into:

D N S

I P

T C P / I P

etc. Lets say I have this sentence

Because IP provides this basic routing function, the term “IP router,” is often used. Other, older terms for router are (IP gateway), [Internet gateway], and 'gateway'. TCP/IP 12345.

Running this command kinda solves my problem: sed -e "s/[a-z \, \. \' \“ \” \( \) 0-9]*/& /g" -e "s/ */ /g" test.txt It doesn't get it quite perfectly.

I get this:

Because I P provides this basic routing function, the term “ I P router,” is often used. Other, older terms for router are ( I P gateway), [ Internet gateway ], and 'gateway'. T C P / I P 12345.

There is still a space between " and I P .

There is a space between ( and I P .

There is also a space between [ and Internet.

escaping [ and ] with $ sed -e "s/[a-z \, \. \' \“ \” \( \) \[ \] 0-9]*/& /g" -e "s/ */ /g" test.txt doesn't work as shown below.

Because IP provides this basic routing function, the term “IP router,” is often used. Other, older terms for router are (IP gateway), [Internet gateway], and 'gateway'. TCP/IP 12345.

sed awk text regex
2 Answers

With GNU sed and a conditional jump:

echo 'think of DNS, IP, TCP/IP, etc.' | sed -E ':x; s/([A-Z/])([A-Z/])/\1 \2/; tx'

Output:

think of D N S, I P, T C P / I P, etc.
user_37768580

The regular expression

/([A-Z])([A-Z])/

Will match instances of two capital letters next to each other. You'd then want to use capturing groups in the replacement to get the same letters with a space between them.

/\1 \2/

This will only catch the first two capital letters in a row, so the output after the first iteration would look like:

Think of D NS, I P, T CP/I P, etc.

Therefore you need to repeat the substitution until the regex doesn't match. In Python this would be:

the_string = 'Think of DNS, IP, TCP/IP, etc.'

while re.search(r'([A-Z])([A-Z])', the_string):
    the_string = re.sub(r'([A-Z])([A-Z])', r'\1 \2', the_string)

the_string now ends up as:

Think of D NS, I P, T CP/I P, etc.
user_175684690
Related