Elasticsearch(038)：es中搜索之入门与分页搜索

最新推荐文章于 2024-07-14 14:31:28 发布

瘦子没有夏天

最新推荐文章于 2024-07-14 14:31:28 发布

阅读量639

点赞数

分类专栏： ElasticSearch 文章标签： es dsl es分页 es深度分页 es分页报错 es最新版

本文链接：https://blog.csdn.net/weixin_39723544/article/details/109995523

版权

ElasticSearch 专栏收录该内容

41 篇文章 47 订阅

订阅专栏

ES中搜索的方式分为两种，一种是通过URL参数进行搜索；另一种是通过POST中body请求参数进行搜索。

针对第一种我们进行简单的讲解，我们关注的核心放在第二种上面。

一、简单搜索(search)

1.1 测试使用的index、mapping及数据准备

PUT example
PUT example/docs/_mapping
{
    "properties": {
        "id": {
            "type": "long"
        },
        "orderNo": {
            "type": "keyword"
        },
        "buyerId": {
            "type": "long"
        },
        "shopId": {
            "type": "long"
        },
        "shopName": {
            "type": "text",
            "analyzer":"ik_max_word",
            "search_analyzer":"ik_smart",
            "fields": {
                "keyword": {
                    "type": "keyword"
                }
            }
        },
        "tradeStatus": {
            "type": "integer"
        },
        "payStatus": {
            "type": "integer"
        },
        "orderAmount": {
            "type": "float"
        },
        "payAmount": {
            "type": "float"
        },
        "payPlatformOrderNo": {
            "type": "keyword"
        },
        "delivery": {
            "properties": {
                "receiver": {
                    "type": "keyword"
                },
                "phoneNumber": {
                    "type": "keyword"
                },
                "address": {
                    "type": "keyword"
                },
                "detailedAddress": {
                    "type": "keyword"
                }
            }
        },
        "items": {
            "type": "nested",
            "properties": {
                "id": {
                    "type": "long"
                },
                "orderId": {
                    "type": "long"
                },
                "goodId": {
                    "type": "long"
                },
                "goodName": {
                    "type": "keyword"
                },
                "goodPrice": {
                    "type": "float"
                },
                "goodNumber": {
                    "type": "integer"
                },
                "goodAmount": {
                    "type": "float"
                },
                "createTime": {
                    "type": "date",
                    "format": "yyyy-MM-dd HH:mm:ss",
                    "ignore_malformed": true
                }
            }
        },
        "createTime": {
            "type": "date",
            "format": "yyyy-MM-dd HH:mm:ss",
            "ignore_malformed": true
        },
        "deliveryTime": {
            "type": "date",
            "format": "yyyy-MM-dd HH:mm:ss",
            "ignore_malformed": false
        },
        "deliveryAmount": {
            "type": "float"
        },
        "remark": {
            "type": "text",
            "analyzer":"ik_max_word",
            "search_analyzer":"ik_smart"
        }
    }
}

#bulk插入
POST example/docs/_bulk
{"index": {"_id": 1}}
{"id":1,"orderNo":1136482742505680898,"buyerId":4500,"shopId":9596,"shopName":"小米手机官方旗舰店","tradeStatus":1,"payStatus":1,"orderAmount":4290.90,"payAmount":4280.00,"payPlatformOrderNo":"9b235e16-ffa2-4605-af8c-47bee94d1732","delivery":{"receiver":"张三","phoneNumber":"13279220508","address":"陕西省西安市雁塔区","detailedAddress":"小寨东路91号陕西历史博物馆"},"items":[{"id":123,"orderId":1,"goodId":54,"goodName":"小米CC9Pro","goodPrice":3299.00,"goodNumber":5468,"goodAmount":1,"createTime":"2019-11-11 00:01:18"},{"id":124,"orderId":1,"goodId":55,"goodName":"小米插线板","goodPrice":49.00,"goodNumber":5469,"goodAmount":2,"createTime":"2019-11-11 00:01:18"},{"id":125,"orderId":1,"goodId":56,"goodName":"小米平衡车","goodPrice":1999.00,"goodNumber":5470,"goodAmount":1,"createTime":"2019-11-11 00:01:18"}],"createTime":"2019-11-11 00:01:18","deliveryTime":"2019-11-11 16:30:22","deliveryAmount":20.00,"remark":"尽快发货，谢谢！"}
{"index": {"_id": 2}}
{"id":2,"orderNo":1136482742505656781,"buyerId":4500,"shopId":9500,"shopName":"华为旗舰店","tradeStatus":1,"payStatus":1,"orderAmount":7999.99,"payAmount":6500.00,"payPlatformOrderNo":"9b235e16-ffa2-4605-af8c-47bee94d2134","delivery":{"receiver":"张三","phoneNumber":"13279220508","address":"陕西省西安市雁塔区","detailedAddress":"小寨东路91号陕西历史博物馆"},"items":[{"id":124,"orderId":2,"goodId":1003,"goodName":"华为笔记本","goodPrice":7999.99,"goodNumber":5400,"goodAmount":1,"createTime":"2019-11-11 00:01:18"}],"createTime":"2019-11-11 00:01:18","deliveryTime":"2019-11-12 18:30:22","deliveryAmount":0.00,"remark":"尽快发货，谢谢！"}
{"index": {"_id": 3}}
{"id":3,"orderNo":1136482742505655433,"buyerId":4501,"shopId":8503,"shopName":"无印良品旗舰店","tradeStatus":1,"payStatus":1,"orderAmount":2036.00,"payAmount":2020.00,"payPlatformOrderNo":"9b235e16-ffa2-4605-af8c-47bee94d2300","delivery":{"receiver":"李四","phoneNumber":"13279228888","address":"陕西省西安市长安区","detailedAddress":"西长安街1号大厦"},"items":[{"id":620,"orderId":3,"goodId":8784,"goodName":"柔和洗面奶","goodPrice":79.90,"goodNumber":8780,"goodAmount":1,"createTime":"2019-11-11 00:21:50"},{"id":622,"orderId":3,"goodId":8704,"goodName":"化妆水","goodPrice":109.90,"goodNumber":8580,"goodAmount":2,"createTime":"2019-11-11 00:21:50"}],"createTime":"2019-11-11 00:21:50","deliveryTime":"2019-11-13 17:35:23","deliveryAmount":8.00,"remark":"尽快发货，谢谢！"}

1.2 示例

在这里我们只做简单的示例，不做过多的学习，有过Http基础的话，基本没有什么难度。

#查询所有
GET example/docs/_search

#带有参数的查询
GET example/docs/_search?

#在索引example和us的所有类型中搜索
GET example,us/_search

# 在以g或u开头的索引的所有类型中搜索
GET e*,u*/_search

#在索引example和us的类型为docs和docs2中搜索
GET example,us/docs,docs2/_search

注意：当你搜索包含单一索引时，Elasticsearch转发搜索请求到这个索引的主分片或每个分片的复制分片上，然后聚集每个分片的结果。搜索包含多个索引也是同样的方式——只不过或有更多的分片被关联。

可以跟的参数如下图所示。其实这些方式归根到底第二种方式的功能是重合的。

在这里插入图片描述

二、结构化搜索(request body search)

请求体的参数是JSON格式。这种JSON必须遵循DSL(domain search language)语法格式。在这里我们简单做一下示例使用。关于更多的查询和使用会在后续的文章逐一详细介绍。

2.1 查询所有

GET example/docs/_search
{
  "query": {
    "match_all": {}
  }
}

2.2 分页搜索

和SQL使用LIMIT关键字返回只有一页的结果一样，Elasticsearch接受from和size参数：

size: 结果数，默认10

from: 跳过开始的结果数，默认0。可理解成元素开始的索引下标。

例如我们想分页搜索订单的第一页，每页显示2个。

GET example/docs/_search
{
  "query": {
    "match_all": {}
  },
  "from": 0,
  "size": 2
}

应该当心分页太深或者一次请求太多的结果。结果在返回前会被排序。但是记住一个搜索请求常常涉及多个分片。每个分片生成自己排好序的结果，它们接着需要集中起来排序以确保整体排序正确。

小贴士

在集群系统中深度分页。为了理解为什么深度分页是有问题的，让我们假设在一个有5个主分片的索引中搜索。当我们请求结果的第一页（结果1到10）时，每个分片产生自己最顶端10个结果然后返回它们给请求节点(requesting node)，它再排序这所有的50个结果以选出顶端的10个结果。

现在假设我们请求第1000页——结果10001到10010。工作方式都相同，不同的是每个分片都必须产生顶端的10010个结果。然后请求节点排序这50050个结果并丢弃50040个！你可以看到在分布式系统中，排序结果的花费随着分页的深入而成倍增长。这也是为什么网络搜索引擎中任何语句不能返回多于10000个结果的原因（通常最大的分页数就是100页，比如说京东的评论数等等）。