请查看我的数组。所以，我需要按产品和尺寸而不是价格来分组。意味着，在尺寸和产品相同的情况下，得到价格的总和，然后用这种格式显示在其他数组中。产品|尺寸|价格。

Array
(
[0] => Array
    (
        [0] => size
        [1] => product
        [2] => date
        [3] => price
    )
[1] => Array
    (
        [0] => Large
        [1] => test2
        [2] => 5/9/2016
        [3] => 14
    )
[2] => Array
    (
        [0] => small
        [1] => test3
        [2] => 5/10/2016
        [3] => 17
    )
[3] => Array
    (
        [0] => Large
        [1] => test2
        [2] => 5/9/2016
        [3] => 17
    )
[4] => Array
    (
        [0] => small
        [1] => test
        [2] => 5/10/2016
        [3] => 1
    )
[5] => Array
    (
        [0] => Large
        [1] => test
        [2] => 5/10/2016
        [3] => 1
    )

我试过这个

$counter = 0;

foreach（array_slice（$top100SitesCSV，1）为$key=>$value）｛

if (isset($topSiteArray[$value[1]]['price']) && isset($topSiteArray[$value[1]]['size']) )
    {
    $topSiteArray[$value[1]]['price'] = $topSiteArray[$value[1]]['price'] + $value[3];
    }
  else
    {
    $topSiteArray[$value[1]]['price'] = $value[3];
    $topSiteArray[$value[1]]['size'] = $value[0];
    }
}

伪代码：

foreach（$initial_array为$tmpArr）｛｝
连接$keyStr=$tmpArr[0]$tmpArr[1]
使用给定的$keyStr将$tmpArr推送到$newArray
对于所有其他$tmpArray-s，您连接0,1，并将其用作ID来检查$newArray中是否有类似的键
如果你有，你只需要增加价格，如果你没有，你就用"ID"字符串推送temparray
完成所有操作后，只需将$newArray的键转换为数字即可

我真的不确定u是否是这样的意思，但下面有一个简单的例子，它应该是一个很好的起点。应该添加一些数据验证，并且应该添加标签级数据表的动态映射。

 $data = [ 
    0 => [0 => 'size', 1 => 'product', 2 => 'date', 3 => 'price'],
    1 => [0 => 'large', 1 => 'test', 2 => '5/9/2016', 3 => '5'],
    2 => [0 => 'small', 1 => 'test1', 2 => '5/9/2016', 3 => '10'],
    3 => [0 => 'large', 1 => 'test', 2 => '5/9/2016', 3 => '13'],
    4 => [0 => 'small', 1 => 'test1', 2 => '5/9/2016', 3 => '22'],
    5 => [0 => 'medium', 1 => 'test3', 2 => '5/9/2016', 3 => '5'],    
    ];

$labels = $data[0]; // extract labels from data table 
unset($data[0]); // delete extracted data
$result = []; // prepare results container
foreach($data as $product){ // loop through all products
    if(!isset($result[$product[1]])){ // if prod isn't in table create new result row
        $result[$product[1]] = ['size' => $product[0], 'item' => 1, 'price' => $product[3]];
    } else { // if it`s just increment item count and calculate total price
        $result[$product[1]]['price'] += $product[3];
        $result[$product[1]]['item'] += 1;
    }
}
var_dump($result); // you can filter out of that table results for less than 2 items, or just don`t save it to the result in above loop

结果

 array(3) {
  ["test"]=>
  array(3) {
    ["size"]=>
    string(5) "large"
    ["item"]=>
    int(2)
    ["price"]=>
    int(18)
  }
  ["test1"]=>
  array(3) {
    ["size"]=>
    string(5) "small"
    ["item"]=>
    int(2)
    ["price"]=>
    int(32)
  }
  ["test3"]=>
  array(3) {
    ["size"]=>
    string(6) "medium"
    ["item"]=>
    int(1)
    ["price"]=>
    string(1) "5"
  }
}

这就是我如何理解您描述不周的请求：

代码

<?php
error_reporting(E_ALL);
$raw = array(
    [ 'size',  'product', 'date',      'price' ],
    [ 'Large', 'test2',   '5/9/2016',       14 ],
    [ 'small', 'test3',   '5/10/2016',      17 ],
    [ 'Large', 'test2',   '5/9/2016',       17 ],
    [ 'small', 'test',    '5/10/2016',       1 ],
    [ 'Large', 'test',    '5/10/2016',       1 ]
);
$keys = array_shift($raw);
$products = array();
foreach ($raw as $item) {
    $item = array_combine($keys, $item);
    $size    = array_shift($item);
    $product = array_shift($item);
    $products[$product][$size][] = $item;
}
var_dump($products);

在3v4l.org 上与我一起玩

输出

array(3) {
  ["test2"]=> array(1) {
    ["Large"]=> array(2) {
      [0]=> array(2) {
        ["date"]=> string(8) "5/9/2016"
        ["price"]=> int(14)
      }
      [1]=> array(2) {
        ["date"]=> string(8) "5/9/2016"
        ["price"]=> int(17)
      }
    }
  }
  ["test3"]=> array(1) {
    ["small"]=> array(1) {
      [0]=> array(2) {
        ["date"]=> string(9) "5/10/2016"
        ["price"]=> int(17)
      }
    }
  }
  ["test"]=> array(2) {
    ["small"]=> array(1) {
      [0]=> array(2) {
        ["date"]=> string(9) "5/10/2016"
        ["price"]=> int(1)
      }
    }
    ["Large"]=> array(1) {
      [0]=> array(2) {
        ["date"]=> string(9) "5/10/2016"
        ["price"]=> int(1)
      }
    }
  }
}

编辑：我有点改变了大小，而不是复制，所以它不在最终的$products数组中。但是您看到了其中的逻辑，并有望自己调整代码。

将数组中两个相同的索引进行比较，得到第三个索引的和

comapare the two same indexes in array and then get the sum of third index

结果