I have a long text that contains data like:

23cm,
23m,
60 cm,
60 m,

So sometimes there is a space between number and unit. Sometimes there isn't one. How to add an underscore in each case, so the result would be:

23_cm,
23_m,
60_cm,
60_m

The search pattern for a part of it is probably (\d) (?:cm|m), but I can figure out the rest.

CodePudding user response：

I suggest replacing matches of

(?<=\d) ?(?=c?m,)

with an underscore. If a space is present it is matched; else the (zero-width) location between the last digit and 'cm' or 'm' is matched.

Demo

The regular expression can be broken down as follows. (I have enclosed the space in a character class to make it visible to the reader.)

(?<=     # begin a positive lookbehind    
  \d     # match a digit
)        # end positive lookbehind
[ ]?     # optionally match a space
(?=      # begin a positive lookahead  
  c?m,   # optionally match a 'c' followed by 'm,'
)        # end positive lookahead

If the comma is not always present replace (?=c?m,) with (?=c?m\b), \b being a word boundary.

CodePudding user response：

We can use capturing groups. The following example uses \2 and \3 for the capturing groups. Some languages would use $2 and $3.
See https://regex101.com/r/KxYyrb/1
input string

23cm, 23m, 60 cm, 60 m,

pattern

((\d )\s?(m|cm))

replace using

\2_\3

output

23_cm, 23_m, 60_cm, 60_m,

CodePudding user response：

You can use

(\d)\s?(c?m)\b

The replacement pattern is $1_$2.

See the regex demo.

Details:

• (\d) - Capturing group 1: a digit
• \s? - an optional whitespace char
• (c?m) - Capturing group 2: an optional c and an m
• \b - a word boundary (else, the regex will match m in men, for example).