extract word and before word and insert between ”_” in regex

Question

I need some help on declaring a regex. My inputs are like the following:

I need to extract word and before word and insert between ”_” in regex:python Input

 Input
 s2 = 'Some other medical terms and stuff diagnosis of R45.2 was entered for  this patient. Where did Doctor Who go? Then xxx feea fdsfd'
 # my regex pattern
 re.sub(r"(?:[a-zA-Z'-]+[^a-zA-Z'-]+){0,1}diagnosis", r"\1_", s2)
 Desired Output:
 s2 = 'Some other medical terms and stuff_diagnosis of R45.2 was entered      for this patient. Where did Doctor Who go? Then xxx feea fdsfd'

You have no capturing group but are using `\1` to refer to it . — Wiktor Stribiżew, Mar 05 '19 at 09:11
Try `re.sub(r"[^\w'-]+(?=diagnosis)", "_", s2)`, see [regex demo](https://regex101.com/r/ugW0q4/1). — Wiktor Stribiżew, Mar 05 '19 at 09:15
What Wiktor tries to say is that groups must be captured example: `(?: text not captured ) ( text captured as \1 )+ (text captured as \2)?`. See https://stackoverflow.com/questions/36524507/notation-in-regular-expression — Lokinou, Mar 05 '19 at 09:16
won't be `str.replace(' diagnosis ','_diagnosis')` much better ? — sahasrara62, Mar 05 '19 at 09:42

score 2 · Accepted Answer · answered Mar 05 '19 at 09:22

You have no capturing group defined in your regex, but are using \1 placeholder (replacement backreference) to refer to it.

You want to replace 1+ special chars other than - and ' before the word diagnosis, thus you may use

re.sub(r"[^\w'-]+(?=diagnosis)", "_", s2)

See this regex demo.

Details

[^\w'-]+ - any non-word char excluding ' and _
(?=diagnosis) - a positive lookahead that does not consume the text (does not add to the match value and thus re.sub does not remove this piece of text) but just requires diagnosis text to appear immediately to the right of the current location.

Or

re.sub(r"[^\w'-]+(diagnosis)", r"_\1", s2)

See this regex demo. Here, [^\w'-]+ also matches those special chars, but (diagnosis) is a capturing group whose text can be referred to using the \1 placeholder from the replacement pattern.

NOTE: If you want to make sure diagnosis is matched as a whole word, use \b around it, \bdiagnosis\b (mind the r raw string literal prefix!).

extract word and before word and insert between ”_” in regex

1 Answers1