在Elasticsearch使用URI花式搜索

通过URI实现搜索

URI Search 是通过在URI上传入参数定义你的查询条件，不支持所有的DSL但是更加便利，常用传入参数如下：

curl -XGET -H "Content-Type:application/json" "http://localhost:9200/<index>/_search?q=<parameter>" 

• default_operator
  设置字符串查询时的操作符，共有三类OR、AND、NOT，在查询时术语term将可能进行分组

curl -XGET -H "Content-Type:application/json" "http://localhost:9200/<index>/_search?/_search?q=customer_full_name:(Eddie  Underwood)&default_operator=AND" -d '{"profile":"true"}'

• q
  这个参数需要填写查询语句，Lucene中的字符串查询语法Query string syntax，在使用时会有指定段查询、泛查询、term分组、语句phrase查询、范围查询、正则表达式、通配符查询（效率低占内存大，通配符在查询语句前方则存在更大的问题）、模糊匹配与近似查询这类概念。

curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=customer_last_name:Riley"

• df
  在q参数的查询语句中没有指定field则使用df参数的值指定q中值所属的参数。

curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=Riley&df=customer_last_name"
# 查询customer_last_name为Riley的文档数据

• explain
true或false是否开启分数计算解释，为查询出来的每一个文档，解释分数计算的过程(IDF计算跟词频相关)，默认为false

curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=customer_last_name:Riley&explain=true"

• from
  文档开始的偏移量，与size合用进行分页，但是也存在深度分页的性能问题，默认为0

curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=customer_last_name:Riley&from=10"

• size
  设置命中(his)的文档返回多少条，默认为10条

curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=customer_last_name:Riley&from=10&size=1"

• _source
  可以设置true or false，则表示是否返回文档中_source这个元数据，也可以是字段集合，则表示需要返回原文哪些字段（field）

curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=customer_last_name:Riley&_source=false"
# 不返回_source字段即原文
curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=customer_last_name:Riley&_source=currency,customer_full_name"
# 只返回currency和customer_full_name两个字段

• _source_excludes
  传入字段名集合，返回文档中需要排除的字段

curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=customer_last_name:Riley&_source_excludes=manufacturer,products"
# 将返回文档中的manufacturer和products两个字段排除掉

• _source_includes
  传入字段名集合，返回文档中只需要的字段

curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=customer_last_name:Riley&_source_includes=customer_full_name"
# 只返回文档中的customer_full_name字段

• analyzer
  设置q参数中的查询语句使用的分词器（中文分词器需要提前安装并重启集群），集群默认使用standard分词器

curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=Mary Bailey&df=customer_last_name&analyzer=simple"
# 此处使用的simple分词器，但是应为Token Filters会将单词转成小写，故有可能查不到

• analyze_wildcard
  默认情况下，不分析查询字符串中的通配符术语。 将此值设置为true，将尽力分析这些值，默认为false

curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=725?4*&df=order_id&analyze_wildcard=false"

• sort

• stored_fields
  返回文档的选择性字段，该字段值存储于原文_source之外，在指定mappings是设置在每个字段

# 设置字段是否在store存储
PUT test_demo
{
  "mappings": {
    "properties": {
      "full_name":{
        "type": "keyword",
        "store": true
      }
    }
  }
}
# 调用API查询full_name字段，命中则只返回full_name字段
curl -XGET 'http://localhost:9200/test_demo/_search?q=customer_last_name:"Mary Bailey"&stored_fields=full_name'

• timeout
  设置超时时间
• search_type
  搜索类型，目前只有query_then_fetch和dfs_query_then_fetch两个可用选项，默认是query_then_fetch
• terminate_after
  为每个分片最大的检索文档的值，到达后将提前终止

curl -XGET "http://localhost:9200/kibana_sample_data_ecommerce/_search?q=elys*&df=user&_source=user&terminate_after=1"
# 最终只命中一条文档

指定字段和泛查询

1. 指定字段
   可以在q参数中指定，或者使用df参数指定查询字段

 GET kibana_sample_data_ecommerce/_search?q=customer_last_name:Riley
{
    "profile": "true" //这个选项打开主要是为能够查看到查询的过程，下列同样如此
}
    
//响应如下，只贴出查询过程部分
{
    "type" : "TermQuery",   //@1
    "description" : "customer_last_name:riley",     //@2
    "time_in_nanos" : 299386,
    "breakdown" : { ... }
}

@1 精确查询term精确查询
@2 精确查询customer_last_name值为riley的文档
2. 泛查询

 GET kibana_sample_data_ecommerce/_search?q=Riley
{
    "profile": "true" //这个选项打开主要是为能够查看到查询的过程，下列同样如此
}
    
//响应如下，只贴出查询过程部分
{
    "type" : "DisjunctionMaxQuery",   //@1
    "description" : "(products.manufacturer:riley | MatchNoDocsQuery("failed [products.base_unit_price] query, caused by number_format_exception:[For input string: "Riley"]") ",     //@2
    "time_in_nanos" : 299386,
    "breakdown" : { ... }
}

@1 使用DisjunctionMaxQuery生成多个子查询的合集查询，是最高的评分作为最终得分
@2 精确查询customer_last_name值为riley的文档

Term与语句(phrase)查询

1. Term术语精准查询
2. 语句phrase查询

GET kibana_sample_data_ecommerce/_search?q="Kamal Richards"&df=customer_full_name
{
  "profile": "true"
}

//响应如下，只贴出查询过程部分
{
    "type" : "PhraseQuery", //@1
    "description" : """customer_full_name:"kamal richards"""", //@2
    "time_in_nanos" : 308390,
    "breakdown" : { ... }
}

@1 使用的PhraseQuery语句查询
@2 指定查询customer_full_name的值为kamal richards，语句查询要保证查询条件的顺序

范围查询和算数符号

1. 范围查询

GET kibana_sample_data_ecommerce/_search?q=taxful_total_price:[30.00 TO 43.0]
{
  "profile": "true"
}

//响应如下，只贴出查询过程部分
{
    "type" : "IndexOrDocValuesQuery", //@1
    "description" : "taxful_total_price:[30.0 TO 43.0]", //@2
    "time_in_nanos" : 286535,
    "breakdown" : {...}
}

2. 算数符号进行范围查询

GET kibana_sample_data_ecommerce/_search?q=order_date:(>=2020-03-27 AND <2020-03-28)
{
  "profile": "true"
}

//
{
    "type" : "BooleanQuery", //@1
    "description" : "+order_date:[1585267200000 TO 9223372036854775807] +order_date:[-9223372036854775808 TO 1585353599999]", //@2
    "time_in_nanos" : 616947,
    "breakdown" :  : {...},
    "children" : [ //@3
        {
            "type" : "IndexOrDocValuesQuery", //@4
            "description" : "order_date:[1585267200000 TO 9223372036854775807]",  //@5
            "time_in_nanos" : 227389,
            "breakdown" : {...}
        },
        {
            "type" : "IndexOrDocValuesQuery",
            "description" : "order_date:[-9223372036854775808 TO 1585353599999]",
            "time_in_nanos" : 137528,
            "breakdown" : {...}
        }
    ]
}

@1 使用BooleanQuery对查询进行分组
@2 对查询的描述，必须order_date满足[1585267200000 TO 9223372036854775807]和[-9223372036854775808 TO 1585353599999]两个条件同时满足
@3 本次查询包含的子查询，即BooleanQuery将条件分成了两个IndexOrDocValuesQuery查询
@4
@5

通配符查询

正则表达式

模糊匹配与近似查询

Request Body Search查询

使用SearchTemplate查询

允许你使用mustache language对现有模板进行填充

GET shakespeare/_search/template

POST kibana_sample_data_ecommerce/_search
{
	"profile": true,
	"query": {
		"match_all": {}
	}
}
GET kibana_sample_data_ecommerce/_search
{
	"profile": true,
	
	"query": {
		"match_phrase": {
		  "FIELD": "PHRASE"
		}
	}
}

GET kibana_sample_data_ecommerce/_search?q=customer_full_name:(Eddie  Underwood)&default_operator=AND
{
	"profile":"true"
}


GET kibana_sample_data_ecommerce/_search/template?pretty
{
  "source": {
    
    "query": {
      "match_phrase": {
        "{{my_field}}": "{{my_value}}"
      }
    },
    
    "size": "{{my_size}}"
  },
  "params": {
    "my_routing":"pVeQmXABN95HgAacawWq",
    "my_field": "customer_full_name",
    "my_value": "Eddie  Underwood",
    "my_size": 5
  }
}


POST /_scripts/my_search_template?pretty
{
  "script": {
    "lang": "mustache",
    "source": {
      "query": {
        "match_phrase": {
          "{{my_field}}": "{{my_value}}"
        }
      }
    }
  }
}

GET _scripts/my_search_template


GET kibana_sample_data_ecommerce/_search/template
{
  "id": "my_search_template",
  "params": {
    "my_field": "customer_full_name",
    "my_value": "Eddie  Underwood"
  }
}


GET _render/template
{
  "source": "{ \"query\": { \"terms\": {{#toJson}}statuses{{/toJson}} }}",
  "params": {
    "statuses" : {
        "status": [ "pending", "published" ]
    }
  }
}
//验证一个搜索模板
GET _render/template
{
  "id": "my_search_template",
  "params": {
    "my_field": "customer_full_name",
    "my_value": "Eddie  Underwood"
  }
}


GET /_search/template?pretty
{
  "id": "my_search_template",
  "params": {
    "my_field": "customer_full_name",
    "my_value": "Eddie  Underwood"
  },
  "explain": true
}