PHP preg_match_all expression

Question

I have virtually no experience of regx, but trying my best.

I have a string like this:

$fString = "Name=Sök,Value=2,Title=Combine me,Options=[Item1=1,Item2=2,Item3=3]";

I want to get an array looking like this:

Array[0] = "Name=Sök"
Array[1] = "Value=2"
Array[2] = "Title=Combine me"
Array[3] = "Options=[Item1=1,Item2=2,Item3=3]"

What I have managed to do so far is:

preg_match_all("/[^,]*[\w\d]*=[^,]*/",$fString,$Data);

But it I can't figure out how to fix the last "Option".

Array ( [0] => Array ( [0] => Name=S�k [1] => Value=2 [2] => Title=Combine me [3] => Options=[Item1=1 [4] => Item2=2 [5] => Item3=3] ) )

...and why is the result an array inside an array?!?

[EDIT]

I guess I need to explain the whole idea of what I'm trying to do here, I'm not sure I'm on the right track any more.

I have created some classes where I store all the "persistent" variables in an array. I have a function that serializes this array so I can be stored in a database.

I know all about the serialize() function, but I'm doing some filtering so I can't use it as it is, and I also prefer to have it more readable for manual editing. This array can have nested arrays within, that needs to be preserved. When I read it all back from the database, the original array must be created again.

I had it all working with the eval() command but stumbled into trouble where I had nested arrays because of the " or ' characters was breaking the main outer string. So this approach was an attempt to serialize everything without nested strings that needed to be preserved.

So if I can solve the nested data with preg_match_all I'm there, otherwise I need to come up with another solution.

I guess the data needs to be escaped as well, such as the , and [ ]

The result is always in an array. $Data outputs always an multidimensional array with several result types. — powtac
– powtac, Commented Jan 11, 2011 at 10:27
Can there be a comma in the values, such as "Title=Combine,me"? If yes, how is that case handled? — Tomalak
– Tomalak, Commented Jan 11, 2011 at 10:52
You'll need a mini parser then. Look at what @SirDarius did. You only need something to look out for the nearest , or [, and descend on [ or return the collected list on ]. — mario
– mario, Commented Jan 11, 2011 at 15:48
After reading your edit, I can say that this use case calls for JSON serialization through the json_encode function. JSON is human-editable and allows array nesting. Plus, it's damn easy to decode to php variables using json_decode. — SirDarius
– SirDarius, Commented Jan 11, 2011 at 16:57

SirDarius · Accepted Answer · 2011-01-11 10:45:03Z

2

Here is a function that will do basically what you need:

function explode_me($str) {
    $a = array();
    $v = "";
    $ignore = false;
    for ($i = 0; $i < strlen($str); $i++) {
        if ($str[$i] == ',' && !$ignore) {
            $a[] = $v;
            $v = "";
        }
        else if ($str[$i] == '[' && !$ignore) {
            $ignore = true;
            $v .= $str[$i];
        }
        else if ($str[$i] == ']' && $ignore) {
            $ignore = false;
            $v .= $str[$i];
        }
        else {
            $v .= $str[$i];
        }
    }
    $a[] = $v;
    return $a;
}

To test it:

$str = "Name=Sök,Value=2,Title=Combine me,Options=[Item1=1,Item2=2,Item3=3]";
$a = explode_me($str);

print_r($a);

which prints:

Array
(
    [0] => Name=Sök
    [1] => Value=2
    [2] => Title=Combine me
    [3] => Options=[Item1=1,Item2=2,Item3=3]
)

answered Jan 11, 2011 at 10:45

SirDarius

43.3k8 gold badges92 silver badges106 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Max Kielland Over a year ago

I used your JSON suggestion, seems to work as a charm so far :D

Tomalak · Accepted Answer · 2011-01-11 10:36:42Z

2

(\w+)=(\[[^\]]+\]|[^,]+)

This breaks down as:

(\w+)        # a word (store in match group 1)
=            # the "=" character
(            # begin match group 2
  \[         #   a "[" character
  [^\]]+     #   anything but "]" character
  \]         #   a "]" character
  |          #   or...
  [^,]+      #   anything but a comma
)            # end match group 1

Apply with preg_match_all():

$fString = "Name=Sök,Value=2,Title=Combine me,Options=[Item1=1,Item2=2,Item3=3]";

$matches = array();
preg_match_all("/(\\w+)=(\\[[^\\]]+\\]|[^,]+)/", $fString, $matches);

Which results in something even more detailed than you wanted to have:

Array
(
    [0] => Array
        (
            [0] => Name=Sök
            [1] => Value=2
            [2] => Title=Combine me
            [3] => Options=[Item1=1,Item2=2,Item3=3]
        )

    [1] => Array
        (
            [0] => Name
            [1] => Value
            [2] => Title
            [3] => Options
        )

    [2] => Array
        (
            [0] => Sök
            [1] => 2
            [2] => Combine me
            [3] => [Item1=1,Item2=2,Item3=3]
        )

)

$result[0] is what you wanted. $result[1] and $result[2] are property names and values separately, which enables you to use them right away instead of making an extra step that splits things like "Options=[Item1=1,Item2=2,Item3=3]" at the correct =.

answered Jan 11, 2011 at 10:36

Tomalak

340k68 gold badges547 silver badges635 bronze badges

2 Comments

Tomalak Over a year ago

@Max: And that would be what? :)

Max Kielland Over a year ago

Well, why the answer is an array with arrays, very handy :D

Or Weinberger · Accepted Answer · 2011-01-11 10:28:47Z

1

If you could change the separators between the items (where it says Item1=1,Item2=2,Item3=3 to something like Item1=1|Item2=2|Item3=3) you could easily use explode(',',$fString) to convert a string to an array.

I can also offer this piece of code that will change the separators, as I have no experience with regex:

$newstr = str_replace(',Item','|Item',$fString);
$newarray = explode(',',$newstr);

$newarray will look like this:

Array[0] = "Name=Sök"
Array[1] = "Value=2"
Array[2] = "Title=Combine me"
Array[3] = "Options=[Item1=1|Item2=2|Item3=3]"

edited Jan 11, 2011 at 10:28

answered Jan 11, 2011 at 10:21

Or Weinberger

7,51223 gold badges73 silver badges117 bronze badges

1 Comment

Max Kielland Over a year ago

I had that in mind too before, but I need to be able to iterate the [...] blocks back in again to unwrap them.

mario · Accepted Answer · 2011-01-11 10:35:36Z

1

This is a problem that lends itself more to parsing than regex extraction. Bout you can separate the special case to make it work:

preg_match_all("/(\w+)=( \w[^,]+ | \[[^\]]+\] )/x", $str, $m);
$things = array_combine($m[1], $m[2]);

Will give you a PHP variable like (but you can access $m[0] for the unparsed strings):

[Name] => Sök
[Title] => Combine me
[Options] => [Item1=1,Item2=2,Item3=3]

You can reapply the function on Options to explode that too.

The trick again is differentiating between \w anything that starts with a letter, and the \[...\] enclosed options. There you have to just make it match ^] all non-closing-brackets, and that's it.

answered Jan 11, 2011 at 10:35

mario

146k20 gold badges243 silver badges293 bronze badges

2 Comments

Max Kielland Over a year ago

Thanks, this was actually the result I was aiming for. Just one problem: When I iterate the [...] parameter (removing the [...] first) "Item1=1,Item2=2,Item3=3" It fails because there is no match for digits. I tried to put in a \d but can't make it work. Furter I need to be able to have nested [...] blocks such as Controls=[Options=[Item1=1,Item2=2],Name=Foo] Maybe this isn't the right approach?!?

mario Over a year ago

@MaxKielland: You don't need \d. The failure above was the +, change it into a \w[^,]*. But as for nested lists, I'm afraid, regular expressions become difficult to manage. You could try to nest another \\[...\\] within though.

mario · Accepted Answer · 2011-01-11 17:47:19Z

1

So, here is another approach. It's a mini parser for nested structures. Adapt the regex if you need escape codes.

function parse(&$s) {
    while (strlen($s) && preg_match("/^(.*?)([=,\[\]])/", $s, $m)) {
        $s = substr($s, 1 + strlen($m[1]));
        switch ($m[2]) {
            case "=":
               $key = $m[1];
               break;
            case ",":
               if (!isset($r[$key])) {
                  $r[$key] = $m[1];
               }
               break;
            case "[":
               $r[$key] = parse($s);
               break;
            case "]":
               return $r;
        }
    }
    if ($s) { $r[$key] = $s; } // remainder
    return $r;
}

answered Jan 11, 2011 at 17:47

mario

146k20 gold badges243 silver badges293 bronze badges

3 Comments

Max Kielland Over a year ago

I will have a look at this and compare with JSON proposed by SirDarius. I also read a bit about the Serializable Interface and it looks like I can do something in this direction...

mario Over a year ago

@MaxKielland: It's actually based on a JSON parser I wrote. And I've tested it with nested lists already. I've compacted it a bit too much for this case, so you will need print statements to debug what's going on. But give it a try on one of your complex examples.

Max Kielland Over a year ago

I will get back, when I have implemented and tested this. For the moment I'm testing the components to be serialized.

Collectives™ on Stack Overflow

PHP preg_match_all expression

5 Answers 5

1 Comment

2 Comments

1 Comment

2 Comments

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

5 Answers 5

1 Comment

2 Comments

1 Comment

2 Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related