Regex - match an optionally repeated pattern of capture groups-CodePudding

Appologies for not knowing exactly how to word this question. There is probably even a better title. I'm open to suggestions.

I have the following subjects:

(Field1 = 'Value1') and (Field2 = 'Value2')

and

(Field1 = 'Value1') and (Field2 = 'Value2') or (Field3 = 'Value3')

I want to match in such a way that I have each thing between the () in groups and each conjunction in a group. So, for the second one, some variation of

0: Field1 = 'Value1'
1: and
2: Field2 = 'Value2'
3: or
4: Field3 = 'Value3'

The good news is, I've got regex that works on the first:

\(([A-Za-z0-9\s\'=] )\) (and|or) \(([A-Za-z0-9\s\'=] )\)

https://regex101.com/r/hMXAXS/1

But (on the second subject) it doesn't match the third "and ()". I need to support arbitrary numbers of groups. I can modify it to just look for "and ()" but then it doesn't match the first group.

How can I tell regex to do this? I either need to "double count" some groups (which is fine) or have some other way of optionally looking for additional patterns and matching them.

Thanks for the help!

PS: I was able to get my application to work with the regex ((and|or) \(([A-Za-z0-9\s\'=] )\)) and then just accepting that the first group would never match and creating application logic to support this. Still, I'd bet there's a better way.

CodePudding user response：

You may use preg_match_all here with the regex pattern (?<=\()(.*?)(?=\))|(?:and|or) as follows:

$input = "(Field1 = 'Value1') and (Field2 = 'Value2') or (Field3 = 'Value3')";
preg_match_all("/(?<=\()(.*?)(?=\))|(?:and|or)/", $input, $matches);
print_r($matches[0]);

This prints:

Array
(
    [0] => Field1 = 'Value1'
    [1] => and
    [2] => Field2 = 'Value2'
    [3] => or
    [4] => Field3 = 'Value3'
)

CodePudding user response：

If you are not worried about fringe cases where delimiting words or paretheses can exist within the parenthetical expressions, then preg_split() generates the desired flat array.

Code: (Demo)

$input = "(Field1 = 'Val and ue1') and (Field2 = 'Valu or e2') or (Field3 = 'Value3')";
var_export(
    preg_split(
        "~^\(|\)$|\) (and|or) \(~",
        $input,
        0,
        PREG_SPLIT_NO_EMPTY|PREG_SPLIT_DELIM_CAPTURE
    )
);

Output:

array (
  0 => 'Field1 = \'Val and ue1\'',
  1 => 'and',
  2 => 'Field2 = \'Valu or e2\'',
  3 => 'or',
  4 => 'Field3 = \'Value3\'',
)

Or simplify the pattern by pre-trimming the outermost parentheses. (Demo)

var_export(preg_split("~\) (and|or) \(~", trim($input, '()'), 0, PREG_SPLIT_DELIM_CAPTURE));

You can also use the continue metacharacter \G to continue matching from the end of the previous match: (Demo) This takes 88 steps versus Tim's pattern which takes 280 steps to parse the string.

$input = "(Field1 = 'Val and ue1') and (Field2 = 'Valu or e2') or (Field3 = 'Value3')";
preg_match_all('~(?:^\(|\G(?!^)(?:\) | \())\K(?:(?:and|or)|[^)] )~', $input, $m);
print_r($m[0]);