PHP模拟SQL的GROUP BY算法

BY JENNER · 2015年1月24日· 阅读次数：25

github地址：https://github.com/huyanping/Zebra-PHP-ArrayGroupBy

packagist地址：https://packagist.org/packages/jenner/array_group_by

为什么使用Zebra-PHP-ArrayGroupBy

在如下场景中，我们总是希望能够在php中使用类似mysql的groupby操作：

• SQL过于复杂，造成数据库运算效率低下
• 从数据库中读取出原始数据，在php中进行运算，增强代码重用率
• 其他非数据库场景的数组归并场景

Zebar-PHP-ArrayGroupBy能够做什么

• 对二维数组进行归并
• 归并的同时，支持对字段进行自定义处理
• 比SQL更灵活的自定义函数，你可以随意编写归并和字段合并函数

示例：

$records = [
    [‘order_date‘ => ‘2014-01-01‘, ‘price‘ => 5],
    [‘order_date‘ => ‘2014-01-02‘, ‘price‘ => 10],
    [‘order_date‘ => ‘2014-01-03‘, ‘price‘ => 20],
    [‘order_date‘ => ‘2015-01-04‘, ‘price‘ => 25],
];

$group_by_fields = [
    ‘order_date‘ => function($value){
            return date(‘Y‘, strtotime($value));
        }
];

$group_by_value = [
    ‘order_date‘ => [
        ‘callback‘ => function($value_array){
                return substr($value_array[0], 0, 4);
            },
        ‘as‘ => ‘year‘
    ],
    ‘price‘ => function($value_array){
            return array_sum($value_array);
        },
];

$grouped = \Jenner\Zebra\ArrayGroupBy::groupBy($records, $group_by_fields, $group_by_value);
print_r($grouped);

结果：

Array
(
    [0] => Array
        (
            [year] => 2014
            [price] => 35
        )

    [1] => Array
        (
            [year] => 2015
            [price] => 25
        )

)

你也可以使用链式方法调用，对数据进行多次汇总，更加灵活：

$records = [
    [‘bill_time‘=>‘2014-01-01 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-01 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-01 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-01 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-01 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-01 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-02 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-02 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-02 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-02 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-02 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-03 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-03 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-03 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-03 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-03 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-04 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
    [‘bill_time‘=>‘2014-01-04 00:00:00‘, ‘price‘=>1, ‘cnt‘=>3,],
];

$group_by_fields = [
    ‘bill_time‘ => function($field){
            return substr($field, 0, 10);
        }
];

$group_by_values = [
    ‘bill_time‘ => function($field_values){
            return substr($field_values[0], 0, 10);
        },
    ‘price‘ => function($field_values){
            return array_sum($field_values);
        },
    ‘cnt‘ => function($field_values){
            return array_sum($field_values);
        }
];

$week_fields = [
    ‘bill_time‘ => function($field){
            return date(‘w‘, strtotime($field));
        }
];

$week_values = [
    ‘bill_time‘ => function($field_values){
            return date(‘w‘, strtotime($field_values[0]));
        },
    ‘price‘ => function($field_values){
            return array_sum($field_values);
        },
    ‘cnt‘ => function($field_values){
            return array_sum($field_values);
        }
];

$grouped = (new \Jenner\Zebra\ArrayGroupBy($records))->groupByField($group_by_fields)->groupByValue($group_by_values)->groupByField($week_fields)->groupByValue($week_values)->get();

print_r($grouped);

举例

• 归并过程中，实现对结果的中值计算
• 归并过程中，对时间字段进行自定义处理，例如归并每5分钟的数据
• 等等

原创文章，转载请注明： 转载自始终不够

本文链接地址: PHP模拟SQL的GROUP
BY算法