Replace multiple occurences of character after zero-length assertion

Question

I would like to replace every _ with a - on lines starting with #| label: using PCRE2 regex within my text editor.

Example:

#| label: my_chunk_label
my_function_name <- function(x)

Should become:

#| label: my-chunk-label
my_function_name <- function(x)

In contrast to .NET regex, where one could substitute (?<=^#\| label: .+)_ with - (regex101 example), PCRE2 does not support infinite lookbehind so the regex is invalid. So far, the only way I found was to repeatedly substitute ^#[^_]+\K_ with - (regex101 example) but I was curious if there is a single-pass solution.

Also an idea to [skip](https://stackoverflow.com/questions/24534782/how-do-skip-or-f-work-on-regex) the other lines: [`^(?!#\| label:).+(*SKIP)(*F)|_`](https://regex101.com/r/5uujbQ/1) — bobble bubble, Aug 23 '23 at 15:00
@bobblebubble interesting and quite simple to understand (as well as efficient). Thanks! — jkd, Aug 23 '23 at 15:35

The fourth bird · Accepted Answer · 2023-08-23T15:08:11.760

3

If you are using pcre, you could make use of \G and \K

Then in the replacement use -

(?:^#\|\h+label:\h+|\G(?!^))[^\r\n_]*\K_

The pattern matches:

(?: Non capture group for the alternatives
- ^#\|\h+label:\h+ Match the pattern that should be at the start of the string, where \h matches a horizontal whitespace character
- | Or
- \G(?!^) Assert the current position at the end of the previous match, not at the start
) Close the non capture group
[^\r\n_]* Match optional characters except for newlines or _
\K Forget what is matched so far
_ Match the underscore

Regex demo

edited Aug 23 '23 at 15:08

answered Aug 23 '23 at 14:25

The fourth bird

154,723
16
55
70

1

Excellent! Thank you for the perfect and fast reply. I had not yet found out about the `\G` anchor. – jkd Aug 23 '23 at 14:39
first `|`should be escaped, otherwise great answer – Nahuel Fouilleul Aug 23 '23 at 14:56
@NahuelFouilleul You are right, I have updated it. – The fourth bird Aug 23 '23 at 15:08

score 2 · Answer 2 · answered Aug 23 '23 at 17:32

An alternative idea is to use PCRE verbs (*SKIP)(*F) to skip the lines that you don't want.

^(?!#\| label:).+(*SKIP)(*F)|_

It's also related to The Trick. On the left side of the alternation lines that don't start with #| label get identified by use of a negative lookahead, matched by .+ and skipped - no replacment wanted in these lines. On the right side matching is done in the remaining lines.

See this demo at regex101 (replace with whatever you like - add variable space if needed)

Replace multiple occurences of character after zero-length assertion

2 Answers2