java正则表达式涉及正向后视和前瞻

Question

I'm trying to figure out a regex for to find <ANYTHING_BUT_WHITSPACE>? OR ?<ANYTHING_BUT_WHITSPACE> and replace the ? with a blank space.

So, '?test test?' should become 'test test'

Below is the regex i came up with; but doesnt seem to work. Any suggestions?

s.replace("(?<=/S)?|?(?=/S)", "");

• (?<=/S)? look for ? with positive look-behind of anything but whitespace (\\S)
• | or
• ?(?=/S) look for ? with positive lookahead of anything but whitespace (\\S)

Answer 1

First of all your regex have some mistakes. You used / instead of \\ . Second thing is escaping the characters.

The regex you are looking for is (?<=\\S)\\?|\\?(?=\\S) and replace with empty string.

Note: For Java use double escapes ie \\\\S and \\\\? .

Regex101 Demo

Answer 2

• First the ? as a literal needs to be escaped since it is special regex character ( \\\\? instead of ? ).

• You should use replaceAll instead of replace to replace using a regex.

• Also make sure to re-assign the return value of replaceAll to a string because strings are immutable in Java.

• The non-whitespace predefined character group is \\S not /S .

Answer 3

I think you need to escape the question mark if used as a literal in regex
expression because it is a metacharacters.

So, it's probably this you need

 (?:
      (?<! \s )
      \?
   |  
      \?
      (?! \s )
 )

However, it may not look intuitive, but using an expression with a leading
assertion greatly slows down the engine.

To get better results, match the literal first then check with an assertion.

 \?
 (?:
      (?! \s )
   |  (?<! \s \? )
 )

Let's compare the relative performance using Benchmark software.

Input

wrgasgsagasf?afbafbadfbadfbadfbafdb
dddd? asfbasbfasfb
?asvgasgasgasgasg

Bench

Regex1:   \?(?:(?!\s)|(?<!\s\?))
Options:  < none >
Completed iterations:   50  /  50     ( x 1000 )
Matches found per iteration:   3
Elapsed Time:    0.15 s,   154.76 ms,   154756 µs


Regex2:   (?:(?<!\s)\?|\?(?!\s))
Options:  < none >
Completed iterations:   50  /  50     ( x 1000 )
Matches found per iteration:   3
Elapsed Time:    0.89 s,   894.83 ms,   894834 µs