PHP：正则表达式匹配完整的匹配括号？

Question

In PHP I have the following string:

 $text = "test 1
          {blabla:database{test}}
          {blabla:testing}
          {option:first{A}.Value}{blabla}{option:second{B}.Value}
          {option:third{C}.Value}{option:fourth{D}}
          {option:fifth}
          test 2
         ";

I need to get all { option ...} out of this string (5 in total in this string). Some have multiple nested brackets in them, and some don't. Some are on the same line, some are not.

I already found this regex:

(\{(?>[^{}]+|(?1))*\})

so the following works fine :

preg_match_all('/(\{(?>[^{}]+|(?1))*\})/imsx', $text, $matches);

The text that's not inside curly brackets is filtered out, but the matches also include the blabla -items, which I don't need.

Is there any way this regex can be changed to only include the option -items?

Answer 1

I modified your initial expression to search for the string '(option:)' appended with non-whitespace characters (\\S*), bounded by curly braces '{}'.

\{(option:)\S*\}

Given your input text, the following entries are matched in regexpal:

test 1

{blabla:database{test}}

{blabla:testing}

{option:first{A}.Value} {option:second{B}.Value}

{option:third{C}.Value}

{option:fourth{D}}

{option:fifth}

test 2

Answer 2

Try this regular expression - it was tested using .NET regular expressions, it may work with PHP as well:

\{option:.*?{\w}.*?}

Please note - I'm assuming that you have only 1 pair of brackets inside, and inside that pair you have only 1 alphanumeric character

Answer 3

如果你在同一级别没有多对括号，这应该有效

/(\{option:(([^{]*(\{(?>[^{}]+|(?4))*\})[^}]*)|([^{}]+))\})/imsx

Answer 4

This problem is far better suited to a proper parser, however you can do it with regex if you really want to.

This should work as long as you're not embedding options inside other options.

preg_match_all(
    '/{option:((?:(?!{option:).)*)}/',
    $text,
    $matches,
    PREG_SET_ORDER
);

Quick explanation.

{option:               // literal "{option:"
  (                    // begin capturing group
    (?:                // don't capture the next bit
      (?!{option:).    // everything NOT literal "{option:"
    )*                 // zero or more times
  )                    // end capture group
}                      // literal closing brace

var_dump ed output with your sample input looks like:

array(5) {
  [0]=>
  array(2) {
    [0]=>
    string(23) "{option:first{A}.Value}"
    [1]=>
    string(14) "first{A}.Value"
  }
  [1]=>
  array(2) {
    [0]=>
    string(24) "{option:second{B}.Value}"
    [1]=>
    string(15) "second{B}.Value"
  }
  [2]=>
  array(2) {
    [0]=>
    string(23) "{option:third{C}.Value}"
    [1]=>
    string(14) "third{C}.Value"
  }
  [3]=>
  array(2) {
    [0]=>
    string(18) "{option:fourth{D}}"
    [1]=>
    string(9) "fourth{D}"
  }
  [4]=>
  array(2) {
    [0]=>
    string(14) "{option:fifth}"
    [1]=>
    string(5) "fifth"
  }
}