PHP Sophisticated String parsing

Question

This may be able to be accomplished with a regular expression but I have no idea. What I am trying to accomplish is being able to parse a string with a given delimiter but when it sees a set of brackets it parses differently. As I am a visual learning let me show you an example of what I am attempting to achieve. (PS this is getting parsed from a url)

Given the string input:

String1,String2(data1,data2,data3),String3,String4

How can I "transform" this string into this array:

{
    "String1": "String1",
    "String2": [
        "data1",
        "data2",
        "data3"
    ],
    "String3": "String3",
    "String4": "String4
}

Formatting doesn't have to be this strict as I'm just attempting to make a simple API for my project.

Obviously things like

array explode ( string $delimiter , string $string [, int $limit = PHP_INT_MAX ] )

Wouldn't work because there are commas inside the brackets as well. I've attempted manual parsing looking at each character at a time but I fear for the performance and it doesn't actually work anyway. I've pasted the gist of my attempt.

https://gist.github.com/Fudge0952/24cb4e6a4ec288a4c492

can you post your input and the desired output so we can better understand what you're trying to achieve? — Vladimir Ramik
– Vladimir Ramik, Commented Mar 7, 2016 at 1:13
Can you use pipes/different delimiter for different levels of string children? e.g. String1,String2(data1|data2|data3),String3,String4 — Vladimir Ramik
– Vladimir Ramik, Commented Mar 7, 2016 at 1:33
I can.. As it is a personal project but it would be nice to keep the delimiter the same for them all. Because then when there are cascaded object() i.e. string(object1,object2(subobject),object3) it would still work instead of needing a new delimiter for each hence limiting the amount of parsing — Tim
– Tim, Commented Mar 7, 2016 at 1:40

kungphu · Accepted Answer · 2016-03-07 01:16:32Z

1

While you could try to split your initial string on commas and ignore anything in parentheses for the first split, this necessarily makes assumptions about what those string values can actually be (possibly requiring escaping/unescaping values depending on what those strings have to contain).

If you have control over the data format, though, it would be far better to just start with JSON. It's well-defined and well-supported.

answered Mar 7, 2016 at 1:16

kungphu

4,8613 gold badges30 silver badges38 bronze badges

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

5 Comments

Tim Over a year ago

I did ponder the thought of JSON as the API is only outputting in JSON anyway.

VolkerK Over a year ago

But why do you want this made-up syntax instead of json?

Tim Over a year ago

Basically the URL was formed like this: http://example.com/api/contact?_fields=id,staff(first,last),salary and then it output a JSON response. I think I will go with the JSON response as it seems like the most well formed even though I really did want a simple syntax for URL construction

Tim Over a year ago

Accepted as it was the earliest answer and they're all good answers!

kungphu Over a year ago

Thanks. I think you'll end up happier and more productive with an established format that doesn't require you to roll your own parser (with all the risk and overhead that comes with it).

VolkerK · Accepted Answer · 2016-03-07 01:40:01Z

You can either build an ad-hoc parser like (mostly untested):

<?php
$p = '!
    [^,\(\)]+  # token: String
    |,         # token: comma
    |\(        # token: open
    |\)        # token: close
!x';
$input = 'String1,String2(data1,data2,data3,data4(a,b,c)),String3,String4';

preg_match_all($p, $input, $m);
// using a norewinditerator, so we can use nested foreach-loops on the same iterator
$it = new NoRewindIterator(
    new ArrayIterator($m[0])
);

var_export( foo( $it ) );

function foo($tokens, $level=0) {
    $result = [];
    $current = null;
    foreach( $tokens as $t ) {
        switch($t) {
            case ')':
                break; // foreach loop
            case '(':
                if ( is_null($current) ) {
                    throw new Exception('moo');
                }
                $tokens->next();
                $result[$current] = foo($tokens, $level+1);
                $current = null;
                break;
            case ',':
                if ( !is_null($current) ) {
                    $result[] = $current;
                    $current = null;
                }
                break;
            default:
                $current = $t;
                break;
        }   
    }
    if ( !is_null($current) ) {
        $result[] = $current;
    }
    return $result;
}

prints

array (
  0 => 'String1',
  'String2' => 
  array (
    0 => 'data1',
    1 => 'data2',
    2 => 'data3',
    'data4' => 
    array (
      0 => 'a',
      1 => 'b',
      2 => 'c',
    ),
  ),
  1 => 'String3',
  2 => 'String4',
)

(but will most certainly fail horribly for not-well-formed strings)

or take a look at lexer/parser generator like e.g. PHP_LexerGenerator and PHP_ParserGenerator.

fusion3k · Accepted Answer · 2023-02-22 17:13:31Z

1

This is a solution with preg_match_all():

$string = 'String1,String2(data1,data2,data3),String3,String4,String5(data4,data5,data6)';

$pattern = '/([^,(]+)(\(([^)]+)\))?/';

preg_match_all( $pattern, $string, $matches );

$result = array();
foreach( $matches[1] as $key => $val )
{
    if( $matches[3][$key] )
    { $add = explode( ',', $matches[3][$key] ); }
    else
    { $add = $val; }
    $result[$val] = $add;
}

$json = json_encode( $result );

3v4l.org demo

Pattern explanation:

([^,(]+)        group 1: any chars except ‘,’ and ‘(’
(\(([^)]+)\))?  group 2: zero or one occurrence of brackets wrapping:
   └──┬──┘
   ┌──┴──┐
   ([^)]+)      group 3: any chars except ‘,’

edited Feb 22, 2023 at 17:13

answered Mar 7, 2016 at 2:11

fusion3k

11.7k4 gold badges27 silver badges47 bronze badges

3 Comments

Tim Over a year ago

This is exactly what I was originally attempting to achieve! One thing I noticed is that it only supports one level of brackets see here: eval.in/531501 is there any way to support an arbitrary number of levels?

fusion3k Over a year ago

mmhhh... I'm not sure at 100%, but I think it is not possible... BTW, surely I'm not the king of regex, maybe somebody else can provide better answer. However, if you can access to same data in JSON format, I strongly suggest you to use it.

Tim Over a year ago

Yeah I doubt I'd ever have multiple layers of brackets anyway and if it didn't make the URL look hideous I'd go for JSON any day lol.. How I wish I could accept 2 answers lol

Collectives™ on Stack Overflow

PHP Sophisticated String parsing

3 Answers 3

5 Comments

Comments

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

5 Comments

Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related