Regex find instance of dash, but not dash 2020-04-08 13:06发布 站内问答 / 前端开发 1178 3 5

Question

I am sooo close. I am trying code a regex expression for Notepad++ to replace a dash with a space, ignoring dashes already with a pre/post space. I realize I could search/replace " - " with "foobarfoo" then search for "-" replacing for " " then converting "foobarfoo" back to " - ", but damnit - I'm trying to learn regex!

Here's my problem:

Adapter - BNC Male to BNC-Female, Right Angle

to

Adapter - BNC Male to BNC Female, Right Angle

(note the disappearing dash in "BNC Female")

The closest I am getting is using this: /(?:[^( )])\-(?:[^( )])/g

but that results with it finding the single letter ahead, the dash, and the single letter following:

Adapter - BNC Male to BNC-Female, Right Angle

WHY is it selecting the pre/post characters? Is this not:

(?:[^( )]) find anything except a space (as a noncapturing group)...

\- ... that follows with a dash ...

(?:[^( )]) ... and is followed by anything except a space(as a noncapturing group)

I get even closer is I replace the first term with (?=[^( )]) but if I change the third term to (?![^( )]) I'm back to where I started - just selecting the dash in between the two spaces. GRRRR.

More samples here at http://regexr.com/444i2

Answer 1

The most readable option for Notepad++ is an alternation based regex with a conditional replacement pattern:

Find What: (\s-\s)|-
Replace With: (?1$1: )

The (\s-\s)|- pattern either captures into Group 1 a whitespace, - and a whitespace, or just matches -. If Group 1 matches, its value is pasted back where it was (=this match is skipped), else the - in other contexts is replaced with a space.

See a demo below:

Another option is using nested lookarounds:

Find What: -(?<!\s-(?!=\s))
Replace With: (a space)

The pattern matches:

• - - a hyphen
• (?<!\s-(?!=\s)) - this negative lookbehind fails the match if its pattern matches the following pattern immediately to the left of the current location (that is, right after -):
  • \s - a whitespace
  • - - a hyphen (this is necessary to make sure we get to the same place where the lookbehind was triggered)
  • (?!=\s) - the next char is a whitespace (this is a lookahead, it is not possible to make it a simple \s as we cannot let the lookbehind move its pattern position matching, i.e. we can't let it try \s-\s before and including - (it would result in a true result).

Answer 2

Use \w(-)\w to replace all hyphens surrounded by alphabetic characters, digits and underscores, or [^ ](-)[^ ] to replace all hyphens surrounded by non-space characters.

Both work fine in my Notepad++ version with all of your examples.

Answer 3

To ignore dashes already with a pre/post space you could use positive lookarounds to assert that what is on the left and on the right are a non whitespace character \S

In the replacement use a space.

(?<=\S)-(?=\S)

Regex demo