记录一次elasticsearch-php工作过程

初始化

$hosts = array('192.168.30.41');
$this->client = \Elasticsearch\ClientBuilder::create()->setHosts($hosts)->build();

新建和设置index

    $params = [
            'index' => 'order',
            'body' => [
                'settings' => [
                    'max_result_window' => 10000000  #由于默认只能读取前10000条数据，这里设置为100w，但是代价就是分页越靠后，效率越低。也可以使用scan解决
                ],
                'mappings' => [
                    'goods' => [
                        '_source' => [
                            'enabled' => true
                        ],
                        'properties' => [
                            'product_code' => [
                                'type'=>'string',
                                'store'=>'yes',
                                'fielddata'=>true,
                                'fields'=>[
                                    'raw'=>[ #由于需要按照这个字段分组统计，且不能进行分词，固这样配置。统计时字段需要写为 product_code.raw
                                        'type'=>'string',
                                        'index'=>'not_analyzed'
                                    ]
                                ]
                            ],
                            'order_id'=>[
                                'fielddata'=>true,
                                'type'=>'string'
                            ],
                            'price'=>[
                                'type'=>'double'
                            ],
                            'num'=>[
                                'type'=>'integer'
                            ],
                            'pay_time'=>[
                                'type'=>'date',
                                'format'=>'yyyy-MM-dd HH:mm:ss'
                            ],
                            'take_province'=>[
                                'type'=>'string',
                                'fielddata'=>true,
                                'store'=>'yes',
                                'fields'=>[
                                    'raw'=>[
                                        'type'=>'string',
                                        'index'=>'not_analyzed'
                                    ]
                                ]
                            ],
                            'buyer_nike'=>[
                                'type'=>'string',
                                'fielddata'=>true
                            ]
                        ]
                    ]
                ]
            ]
        ];
        $response = $this->client->indices()->create($params);

插入数据(这里引用了官方文档的例子，大数据导入不使用insert，而使用更为效率的bulk)

$params = ['body' => []];

for ($i = 1; $i <= 1234567; $i++) {
    $params['body'][] = [
        'index' => [
            '_index' => 'my_index',
            '_type' => 'my_type',
            '_id' => $i
        ]
    ];

    $params['body'][] = [
        'my_field' => 'my_value',
        'second_field' => 'some more values'
    ];

    // Every 1000 documents stop and send the bulk request
    if ($i % 1000 == 0) {
        $responses = $client->bulk($params);

        // erase the old bulk request
        $params = ['body' => []];

        // unset the bulk response when you are done to save memory
        unset($responses);
    }
}

// Send the last batch if it exists
if (!empty($params['body'])) {
    $responses = $client->bulk($params);
}

相关查询

1、查询某商品某时间段内订单数、售卖总数和总价格

#where product_code="xxx" and pay_time BETWEEN "2017-01-01 00:00:00" AND "2017-01-31 23:59:59"
$params = [
    'index' => 'order',
    'type' => 'goods',
    'body' => [
        'size' => 1,
        'query' => [
            "bool"=>[
                "must"=>[
                    "term"=>["product_code.raw"=>$code] #上面解释过了，这里采用不分词的统计，使用字段.raw
                ],
                "filter"=>[
                    "range"=>[
                        "pay_time"=>[
                            "gte"=>$start_time,
                            "lte"=>$end_time
                        ]
                    ]
                ]
            ]
        ],
        'aggs' => [
            'sum_this_product'=>['sum'=>['field'=>"num"]], #售卖总数量，sum累加
            'total_price'=>['sum'=>['field'=>"price"]],   #总价格
            'distinct_orderid'=>['cardinality'=>['field'=>'order_id']] #去重订单数
        ]
    ]
];
$response = $this->client->search($params);     

2、统计某时间段所有商品的订单数、售卖总数和总价格

#where pay_time BETWEEN "2017-01-01 00:00:00" AND "2017-01-31 23:59:59"
$params = [
    'index' => 'order',
    'type' => 'goods',
    'body' => [
        'size' => 0,
        'query' => [
            "bool"=>[
                "filter"=>[
                    "range"=>[
                        "pay_time"=>[
                            "gte"=>$start_time,
                            "lte"=>$end_time
                        ]
                    ]
                ]
            ]
        ],
        'aggs' => [
            'num'=>[
                'terms'=>[
                    'field'=>'product_code.raw',
                    'size'=>100,
                    'order'=>['sum_this_product'=>'desc'] #根据统计出来的售卖总数排序
                ],
                'aggs'=>[
                    'sum_this_product'=>['sum'=>['field'=>'num']],
                    'total_this_product'=>['sum'=>['field'=>'price']],
                    'distinct_orderid'=>['cardinality'=>['field'=>'order_id']]
                ]
            ]
        ]
    ]
];
$response = $this->client->search($params);

唠叨：

1、这次使用的是docker环境，使用阿里镜像：https://dev.aliyun.com/detail.html?spm=5176.1972343.2.21.F0KOV2&repoId=1209

2、官方文档：https://www.elastic.co/guide/en/elasticsearch/client/php-api/current/index.html

3、本次工作数据量大约1500w，需要复杂的统计和展现，mysql已经不能满足，故使用es。但是es不支持类似mysql：select in select这样的子查询，着实折腾了不少时间

4、感谢一位大神的博客：https://segmentfault.com/a/1190000004433446，这是个文章系列，很值得参考。