asked on

RegEx: Split by non-word char, except inside quotes

Hi,
Been banging my head on this one for a little while. I'm using boost::regex, and I'd like to split a string by non-word characters, except where the non-word character is inside quotes.

An example string:

The gopher's bike wasn't "hot enough" for the judges.

Would get split into:

The
gopher
'
s
bike
wasn
'
t
"hot enough"
for
the
judges
.
The really tricky part is where the quoted string has an escaped quote. For example:

Here is a "string with \"a quote\" inside" of it.

That should be split as

Here
is
a
"string with \"a quote\" inside"
of
it
.
I think boost::regex is Perl compatible, so it shouldn't matter if I'm using boost::regex, or PHP's preg_split, or any other Perl compatible regex engine.

Can anyone offer any suggestions?

P.S. Yes, I'm trying to keep the quotes in the match, as the split examples above show.

mrjoltcola

You can try negative lookbehind to only match non-word patterns that are not prefixed by an escape character (\) but all that will do is make it skip \" and it will still split on the next space inside the quoted string (a[space]quote), so it still won't treat the whole quoted string atomically. You are really asking too much for a simple regex because it needs unlimited lookbehind to do this.

This is really a job for a multi-state lexer, or a recursive parser. A single regex doesn't have enough context to handle all of the possibilities properly. If using lex or flex we can push/pop states when we see certain delimiters, and then treat the characters differently while in that state, but you can't do that with a simple one-line regex.

Usually I don't want to bring another tool into the mix so I approach this type of pattern by writing a simple parser by hand to properly tokenize the quoted strings.

ddrudik

You should consider using a match operation instead of split (the characters matched on would depend on your specific requirements):

Raw Match Pattern:
"[^"]*"|[A-Za-z]+|[^A-Za-z ]
 
$matches Array:
(
    [0] => Array
        (
            [0] => The
            [1] => gopher
            [2] => '
            [3] => s
            [4] => bike
            [5] => wasn
            [6] => '
            [7] => t
            [8] => "hot enough"
            [9] => for
            [10] => the
            [11] => judges
            [12] => .
        )
 
)

Open in new window

ASKER CERTIFIED SOLUTION

ddrudik

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

headzoo

ASKER

Works like a charm. Of course the thing I dislike about those crazy character laden regex's, is it's difficult for me to understand how it works. But it works, so I guess that's good enough. :)

ddrudik

Thanks for the question and the points.

In case you would like more info on the pattern:

The regular expression:
 
(?-imsx:"(?:(?!(?<!\\)")[\S\s])*"|[A-Za-z]+|[^A-Za-z ])
 
matches as follows:
  
NODE                     EXPLANATION
----------------------------------------------------------------------
(?-imsx:                 group, but do not capture (case-sensitive)
                         (with ^ and $ matching normally) (with . not
                         matching \n) (matching whitespace and #
                         normally):
----------------------------------------------------------------------
  "                        '"'
----------------------------------------------------------------------
  (?:                      group, but do not capture (0 or more times
                           (matching the most amount possible)):
----------------------------------------------------------------------
    (?!                      look ahead to see if there is not:
----------------------------------------------------------------------
      (?<!                     look behind to see if there is not:
----------------------------------------------------------------------
        \\                       '\'
----------------------------------------------------------------------
      )                        end of look-behind
----------------------------------------------------------------------
      "                        '"'
----------------------------------------------------------------------
    )                        end of look-ahead
----------------------------------------------------------------------
    [\S\s]                   any character of: non-whitespace (all
                             but \n, \r, \t, \f, and " "), whitespace
                             (\n, \r, \t, \f, and " ")
----------------------------------------------------------------------
  )*                       end of grouping
----------------------------------------------------------------------
  "                        '"'
----------------------------------------------------------------------
 |                        OR
----------------------------------------------------------------------
  [A-Za-z]+                any character of: 'A' to 'Z', 'a' to 'z'
                           (1 or more times (matching the most amount
                           possible))
----------------------------------------------------------------------
 |                        OR
----------------------------------------------------------------------
  [^A-Za-z ]               any character except: 'A' to 'Z', 'a' to
                           'z', ' '
----------------------------------------------------------------------
)                        end of grouping
----------------------------------------------------------------------

Open in new window