我有一个多维数组。

$count = count($main_array); // 6000+

主数组中的每个元素大约有 25 个键

$count_sub_array = count($main_array[0]); // 25

其中一个键有大约 3000 个唯一值。即

for($i=0; $i< $count; $i++){
    if($main_array[$i][$match]){
         array_push($unique_array,$main_array[$i][$match]);
    }
}
$unique_array = array_unique($unique_array); // count - 3000

我想遍历数组中的每个元素并按唯一类型对它们进行分组。而且我还希望根据特定索引之一对元素进行分组，例如

$unique_index = array('2012-08','2012-07','2012-06','2012-05','2012-04','2012-03','2012-02', '2012-01');

#目前唯一索引的个数是13

所以结果看起来像:

array (size=2689)
 0 => 
array (size=5)
  'type' => string 'abc' (length=19)
  'totalCost' => 
    array (size=13)
      '2012-07-01' => float 790.08
      '2012-08-01' => float 1501.84
      '2012-09-01' => float 1568.9
      '2012-10-01' => float 2756.04
      '2012-11-01' => float 2428.42
      '2012-12-01' => float 1901.09
      '2013-01-01' => float 2538.59
      '2013-02-01' => float 1537.57
      '2013-03-01' => float 1674.51
      '2013-04-01' => float 1141.01
      '2013-05-01' => float 764.24
      '2013-06-01' => float 1179.84
      '2013-07-01' => float 1252.66
   'numOrders' => 
    array (size=13)
      '2012-07-01' => int 16
      '2012-08-01' => int 66
      '2012-09-01' => int 65
      '2012-10-01' => int 59
      '2012-11-01' => int 60
      '2012-12-01' => int 47
      '2013-01-01' => int 85
      '2013-02-01' => int 50
      '2013-03-01' => int 46
      '2013-04-01' => int 41
      '2013-05-01' => int 22
      '2013-06-01' => int 32
      '2013-07-01' => int 47
  'TotalRevenue' => 
    array (size=13)
      '2012-07-01' => float 1695.99
      '2012-08-01' => float 7418.17
      '2012-09-01' => float 7827.99
      '2012-10-01' => float 7929.63
      '2012-11-01' => float 4815.74
      '2012-12-01' => float 5592.59
      '2013-01-01' => float 6874.02
      '2013-02-01' => float 11599.49
      '2013-03-01' => float 10358.57
      '2013-04-01' => float 6909.55
      '2013-05-01' => float 6983.38
      '2013-06-01' => float 7211.84
      '2013-07-01' => float 10422.59
  'profit' => 
    array (size=13)
      '2012-07-01' => float 905.91
      '2012-08-01' => float 5916.33
      '2012-09-01' => float 6259.09
      '2012-10-01' => float 5173.59
      '2012-11-01' => float 2387.32
      '2012-12-01' => float 3691.5
      '2013-01-01' => float 4335.43
      '2013-02-01' => float 10061.92
      '2013-03-01' => float 8684.06
      '2013-04-01' => float 5768.54
      '2013-05-01' => float 6219.14
      '2013-06-01' => float 6032
      '2013-07-01' => float 9169.93
1 => 
array (size=5)
  'type' => string 'bcd' (length=26)
  'totalCost' => 
    array (size=13)
      '2012-07-01' => float 599.53
      '2012-08-01' => float 419.18
      '2012-09-01' => float 212.63
      '2012-10-01' => float 462.55
      '2012-11-01' => float 450.08
      '2012-12-01' => float 447.1
      '2013-01-01' => float 657
      '2013-02-01' => float 734.68
      '2013-03-01' => float 1006.91
      '2013-04-01' => float 1497.95
      '2013-05-01' => float 2100.96
      '2013-06-01' => float 3845.33
      '2013-07-01' => float 6817.38
  'numOrders' => 
    array (size=13)
      '2012-07-01' => int 11
      '2012-08-01' => int 33
      '2012-09-01' => int 25
      '2012-10-01' => int 28
      '2012-11-01' => int 33
      '2012-12-01' => int 28
      '2013-01-01' => int 33
      '2013-02-01' => int 45
      '2013-03-01' => int 54
      '2013-04-01' => int 53
      '2013-05-01' => int 69
      '2013-06-01' => int 91
      '2013-07-01' => int 180
  'TotalRevenue' => 
    array (size=13)
      '2012-07-01' => float 1201.32
      '2012-08-01' => float 4889.65
      '2012-09-01' => float 2171.5
      '2012-10-01' => float 2573.75
      '2012-11-01' => float 2375.1
      '2012-12-01' => float 2807.28
      '2013-01-01' => float 2933.05
      '2013-02-01' => float 5534.61
      '2013-03-01' => float 6946.02
      '2013-04-01' => float 8555.78
      '2013-05-01' => float 8202.32
      '2013-06-01' => float 14140.04
      '2013-07-01' => float 22043.54
  'profit' => 
    array (size=13)
      '2012-07-01' => float 601.79
      '2012-08-01' => float 4470.47
      '2012-09-01' => float 1958.87
      '2012-10-01' => float 2111.2
      '2012-11-01' => float 1925.02
      '2012-12-01' => float 2360.18
      '2013-01-01' => float 2276.05
      '2013-02-01' => float 4799.93
      '2013-03-01' => float 5939.11
      '2013-04-01' => float 7057.83
      '2013-05-01' => float 6101.36
      '2013-06-01' => float 10294.71
      '2013-07-01' => float 15226.16
 );

如何优化循环？

目前我有以下逻辑:

 1. get the main array from mysql
 2. get unique elements
 3. foreach unique element
       a. foreach key like '2013-06-01', '2013-07-01'
       b. foreach main array
              match above key i.e. if($time == '2013-06-01')
       c. get $value['totalCost'], $value['numOrders'], $value['profit'], value['revenue']

按照我的说法，这个循环的复杂度是:3000 * 13 * 6000；

提前感谢您的帮助。

最佳答案

无论如何你都需要循环 3 次，但要选择好你的循环。

因为主数组是最大的，所以你只想循环一次。不是先构建日期数组，而是在遍历主数组时构建并填充它。

1. 从mysql获取主数组

2. 只循环一次

   $dates = array();
   $measures = array(
       'totalCost'    => 0,
       'numOrders'    => 0,
       'TotalRevenue' => 0,
       'profit'       => 0
   );
   foreach ($main as $main_element) {
       foreach (array_keys($measures) as $measure) {
           foreach ($main_element[$measure] as $date => $value) {
               if(!isset($dates[$date])) {
                   $dates[$date] = $measures;
               }
               $dates[$date][$measure] += $value;
           }
       }
   }
   

注意:根据您对结果结构的偏好，您可以替换

$dates[$date][$measure]

通过

$dates[$measure][$date]

你还有 3 个循环，但你只循环了 6000 * 4 * 13次，好多了。

关于php多维数组循环优化，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/18131141/