PHP - Nested array keys based on string lines

Question

I need a help. I'm looking for a solution for this problem! I have a file which contains the following pattern:

Brazil|SaoPaulo|Diadema|RuadaFe Brazil|SaoPaulo|Diadema|RuadoLimoeiro Brazil|SaoPaulo|SaoCaetano|RuadasLaranjeiras Brazil|Parana|Curitiba|ComendadorAraujo USA|NewJersey|JerseyCity|WhashingtonBoulervard USA|NewJersey|JerseyCity|RiverCourt

Which should bring after some array key implementation, something like this (after apply json_encode call on php):

{
    "name": "Brazil",
    "children": [
    {
        "name": "SaoPaulo",
        "children": [
            {
                "name": "Diadema",
                "children": [
                    {"name": "RuadaFe"},
                    {"name": "RuadoLimoeiro"}
                ]
            },
            {
                "name": "SaoCaetano",
                "children": [
                    {"name": "RuadasLaranjeiras"}
                ]
            },
        ]
        "name": "Parana",
        "children": [
            {
                "name": "Curitiba",
                "children": [
                    {"name": "ComendadorAraujo"}
                ]
            }
       ]
    },
    "name":"USA",
    "children":[
      {
         "name": "NewJersey",
          "children": [
            {
                "name": "JerseyCity",
                "children": [
                    {"name": "WhashingonBoulevard"},
                    {"name": "RiverCourt"}
                ]
            }
          ]
       }
     ]
  }
 ]

And keep going and going (and even more deeper too). Please, help me team... thanks in advance.

Here what I get until now:

Array ( [Brazil] => Array ( [SaoPaulo] => Array ( [Diadema] => Array ( [RuadoLimoeiro] => )

                [SaoCaetano] => Array
                    (
                        [RuadasLaranjeiras] => 
                    )

            )

        [Parana] => Array
            (
                [Curitiba] => Array
                    (
                        [ComendadorAraujo] => 
                    )

            )

    )

[USA] => Array
    (
        [NewJersey] => Array
            (
                [JerseyCity] => Array
                    (
                        [WhashingtonBoulervard] => 
                        [RiverCourt] => 
                    )

            )

    )

)

And here is the json encoded:

{
"Brazil":{
"SaoPaulo":
   {"Diadema":
       {"RuadoLimoeiro":null},
       "SaoCaetano":{"RuadasLaranjeiras":null}
       },
"Parana":
   {"Curitiba":
       {"ComendadorAraujo":null}
   }
},
"USA":{
"NewJersey":{
"JerseyCity":{
     "WhashingtonBoulervard":null,
     "RiverCourt":null}
      }
    }
}

As you can see, the "names" and "child" is missing because is not an array key structure, also something is wrong, because I'm missing some values on SaoPaulo.

Here is the function:

foreach($strings as $string) {
    $parts = array_filter(explode('|', $string));
    $ref = &$result;
    foreach($parts as $p) {
       // echo $p;
        if(!isset($ref[$p])) {
            $ref[$p] = array();
           // $ref[$p] = array("name"=>$p);
        }

        $ref = &$ref[$p];

    }
    $ref = null;
}

-------------------------------- AFTER SOME ANSWERS --------------------------

{
    "name": "Brazil(country)",
    "children": [
    {
        "name": "SaoPaulo(state)", // only one state
        "children": [
            {
                "name": "Diadema(city)", // only one city
                "children": [
                    {"name": "RuadaFe(street)"}, // two streets under the same city...
                    {"name": "RuadoLimoeiro(street)"}
                ]
            },
            {
                "name": "SaoCaetano(city)",
                "children": [
                    {"name": "RuadasLaranjeiras(street)"}
                ]
            },
        ]
        "name": "Parana(state)",
        "children": [
            {
                "name": "Curitiba(city)",
                "children": [
                    {"name": "ComendadorAraujo(street)"}
                ]
            }
       ]
    },...

I put on parentesis the structure (country, state, city, street) just to clarify what i want. Got it?

Answer 1

It is not outright trivial, but what I did is actually the same as you (the same way as in your edited question), but I also keep track of those arrays to rename the keys. And that's what I do afterwards then:

$separator = [' ', '|'];

$buffer = file_get_contents($file);
$entries = explode($separator[0], $buffer);
$result = [];
$named[] = &$result;
########
foreach ($entries as $entry) {
    $each     = explode($separator[1], $entry);
    $pointer  = & $result;
    while ($current = array_shift($each)) {
        if (!isset($pointer[$current])) {
            unset($children);
            $children = [];
            $named[] = &$children;
            ########
            $pointer[$current] = ['name' => $current, 'children' => &$children];
        }
        $pointer = & $pointer[$current]['children'];
    }
}

foreach($named as $offset => $namedArray) {
    $keys = array_keys($namedArray);
    foreach($keys as $key) {
        $named[$offset][] = &$namedArray[$key];
        unset($named[$offset][$key]);
    }
}

print_r($result);

See it in action.

You probably might want to modify this by checking in the later foreach if children are empty (leaf-nodes) and then removing the children entry as well.

Here the modified foreach at the end and json output:

foreach($named as $offset => $namedArray) {
    $keys = array_keys($namedArray);
    foreach($keys as $key) {
        if (!$namedArray[$key]['children']) {
            unset($namedArray[$key]['children']);
        }
        $named[$offset][] = &$namedArray[$key];
        unset($named[$offset][$key]);
    }
}

echo json_encode($result, JSON_PRETTY_PRINT);

Output:

[
    {
        "name": "Brazil",
        "children": [
            {
                "name": "SaoPaulo",
                "children": [
                    {
                        "name": "Diadema",
                        "children": [
                            {
                                "name": "RuadaFe"
                            },
                            {
                                "name": "RuadoLimoeiro"
                            }
                        ]
                    },
                    {
                        "name": "SaoCaetano",
                        "children": [
                            {
                                "name": "RuadasLaranjeiras"
                            }
                        ]
                    }
                ]
            },
            {
                "name": "Parana",
                "children": [
                    {
                        "name": "Curitiba",
                        "children": [
                            {
                                "name": "ComendadorAraujo"
                            }
                        ]
                    }
                ]
            }
        ]
    },
    {
        "name": "USA",
        "children": [
            {
                "name": "NewJersey",
                "children": [
                    {
                        "name": "JerseyCity",
                        "children": [
                            {
                                "name": "WhashingtonBoulervard"
                            },
                            {
                                "name": "RiverCourt"
                            }
                        ]
                    }
                ]
            }
        ]
    }
]

The full code-example at a glance:

<?php
/**
 * PHP - Nested array keys based on string lines
 * @link http://stackoverflow.com/a/16305236/2261774
 */
$file = 'data://text/plain;base64,' . base64_encode('Brazil|SaoPaulo|Diadema|RuadaFe Brazil|SaoPaulo|Diadema|RuadoLimoeiro Brazil|SaoPaulo|SaoCaetano|RuadasLaranjeiras Brazil|Parana|Curitiba|ComendadorAraujo USA|NewJersey|JerseyCity|WhashingtonBoulervard USA|NewJersey|JerseyCity|RiverCourt');

$separator = [' ', '|'];

$buffer = file_get_contents($file);
$entries = explode($separator[0], $buffer);
$result = [];
$named[] = &$result;
foreach ($entries as $entry) {
    $each     = explode($separator[1], $entry);
    $pointer  = & $result;
    while ($current = array_shift($each)) {
        if (!isset($pointer[$current])) {
            unset($children);
            $children = [];
            $named[] = &$children;
            $pointer[$current] = ['name' => $current, 'children' => &$children];
        }
        $pointer = & $pointer[$current]['children'];
    }
}
unset($pointer);

foreach($named as $offset => $namedArray) {
    $keys = array_keys($namedArray);
    foreach($keys as $key) {
        if (!$namedArray[$key]['children']) {
            unset($namedArray[$key]['children']);
        }
        $named[$offset][] = &$namedArray[$key];
        unset($named[$offset][$key]);
    }
}
unset($named);

echo json_encode($result, JSON_PRETTY_PRINT);

---

your question JSON:

{
    "name": "Brazil",
    "children": [
    {
        "name": "SaoPaulo",
        "children": [
            {
                "name": "Diadema",
                "children": [
                    {"name": "RuadaFe"},
                    {"name": "RuadoLimoeiro"}
                ]
            },

My Answer JSON:

[
    {
        "name": "Brazil",
        "children": [
            {
                "name": "SaoPaulo",
                "children": [
                    {
                        "name": "Diadema",
                        "children": [
                            {
                                "name": "RuadaFe"
                            },
                            {
                                "name": "RuadoLimoeiro"
                            }
                        ]
                    },

The only difference I can spot is that the root-nodes are wrapped inside an array in my answer, but it should be trivial to fetch them out there if that really is your issue ;)

Answer 2

Does it solves your problem? Use $inputString to put your real string.

    <?php

// ------------------------------------------------------ your input goes here ------------------------------------------------------
$inputString = 'Brazil|SaoPaulo|Diadema|RuadaFe Brazil|SaoPaulo|Diadema|RuadoLimoeiro Brazil|SaoPaulo|SaoCaetano|RuadasLaranjeiras Brazil|Parana|Curitiba|ComendadorAraujo USA|NewJersey|JerseyCity|WhashingtonBoulervard USA|NewJersey|JerseyCity|RiverCourt';

class item {
    public $name = null;

    public function getChildrenByName($strName) {

        $ret = null;
        # this variable should be defined in interface, but i skipped it so it wont be printed in json when obj does not have childrens
        if( !isset( $this->children ) ) {
            $this->children = array(  );
        }

        foreach ( $this->children as $child ) {
            if( $child->name === $strName ) {
                $ret = $child;
                break;
            }
        }


        if ( !$ret ) {
            $this->children[] = self::spawnByName( $strName );
        }

        return $this->children[ count($this->children) - 1];
    }

    static public function spawnByName($strName) {
        $ret = new item();
        $ret->name = $strName;
        return $ret;
    }
}

class listManager {

    protected $list = array();

    public function getList() {
        return $this->list;
    }

    public function addPath( $desiredPath ) {

        # path needs to be as array
        if ( is_string( $desiredPath ) ) {
            $desiredPath = explode('|', $desiredPath);
        }

        # create root element if it does not already exists
        if ( !isset( $this->list[$desiredPath[0]] ) ) {
            $this->list[$desiredPath[0]] = item::spawnByName($desiredPath[0]);
        }

        $curElement = $this->list[$desiredPath[0]];

        for( $i=1; $i<count($desiredPath); $i++ ) {
            $curElement = $curElement->getChildrenByName( $desiredPath[$i] );
        }

    }

    protected function spawnElement( $strName ) {
        $ret = new item();
        $ret->name = $strName;
        return $ret;
    }
}

$output = array();
$expl = explode(' ', $inputString);
$list = new listManager();

foreach ( $expl as $key => $path ) {
    $list->addPath( $path );

}

$output = '';

foreach ( $list->getList() as $singleVariable ) {
    $output .= json_encode($singleVariable, JSON_PRETTY_PRINT) . ",\n";
}

echo '<pre>'.$output.'</pre>';

?>

Above code produces following json out of your sample code:

{
    "name": "Brazil",
    "children": [{
            "name": "SaoPaulo",
            "children": [{
                    "name": "Diadema",
                    "children": [{
                            "name": "RuadaFe"
                        }
                    ]
                }
            ]
        }, {
            "name": "SaoPaulo",
            "children": [{
                    "name": "Diadema",
                    "children": [{
                            "name": "RuadoLimoeiro"
                        }
                    ]
                }
            ]
        }, {
            "name": "SaoPaulo",
            "children": [{
                    "name": "SaoCaetano",
                    "children": [{
                            "name": "RuadasLaranjeiras"
                        }
                    ]
                }
            ]
        }, {
            "name": "Parana",
            "children": [{
                    "name": "Curitiba",
                    "children": [{
                            "name": "ComendadorAraujo"
                        }
                    ]
                }
            ]
        }
    ]
} {
    "name": "USA",
    "children": [{
            "name": "NewJersey",
            "children": [{
                    "name": "JerseyCity",
                    "children": [{
                            "name": "WhashingtonBoulervard"
                        }
                    ]
                }
            ]
        }, {
            "name": "NewJersey",
            "children": [{
                    "name": "JerseyCity",
                    "children": [{
                            "name": "RiverCourt"
                        }
                    ]
                }
            ]
        }
    ]
}

edit: changed, does it fit now?