天天看点

Elasticsearch --- (十三)索引管理1、创建、修改以及删除索引2、 修改分词器以及定义自己的分词器3、深入探秘type底层数据结构4、mapping root object深入剖析5、定制化自己的dynamic mapping策略6、基于scroll滚动搜索和alias别名实现zero downtime reindex

目录

1、创建、修改以及删除索引

(1)手动创建索引语法

(2)修改索引

(3)删除索引

2、 修改分词器以及定义自己的分词器

(1)默认分词器   standard

(2)修改分词器的设置

(3)定制化自己的分词器

3、深入探秘type底层数据结构

(1)理论知识

(2)案例实践

(3)最后总结

4、mapping root object深入剖析

(1)root object:某个type对应的mapping json

(2)properties:包含type、index、analyzer

(3)_source

(4)_all

(5)标识性metadata:_index、_type、_id

5、定制化自己的dynamic mapping策略

(1)定制dynamic策略

(2)定制dynamic mapping策略

a、date_detection

b、定制自己的动态映射模板dynamic mapping template(type level)

c、定制自己的default mapping template(index level)

6、基于scroll滚动搜索和alias别名实现zero downtime(零停机) reindex

(1)重建索引

1、创建、修改以及删除索引

(1)手动创建索引语法

PUT /my_index

{

  "settings": {... any settings...},

  "mappings": {

    "type_one":{... any mappings...},

    "type_two":{... any mappings...},

    ...

  }

}

PUT /my_index
{
  "settings": {
    "number_of_shards": 1,
    "number_of_replicas": 0
  },
  "mappings": {
    "my_type":{
      "properties": {
        "my_field":{
          "type":"text"
        }
      }
    }
  }
}
           

(2)修改索引

//---------修改replica 数量
PUT /my_index/_settings
{
  "number_of_replicas": 1
}
           

(3)删除索引

  • DELETE  /my_index
  • DELETE  /index_one,index_two
  • DELETE  /index_*
  • DELETE  /_all             -----删除所有索引
  • elasticsearch.yml 中配置action.destructive_requires_name:true,就不能用 DELETE /_all 这种方式删除全部索引

2、 修改分词器以及定义自己的分词器

(1)默认分词器   standard

  • standard  tokenizer:以单词边界进行切分
  • standard  token  filter:什么都不做
  • lowercase  token  filter:将所有字母转换为小写
  • stop token filter(默认被禁用):移除停用词,比如a    the    it 等等

(2)修改分词器的设置

启用english 停用词 token filter

PUT /my_index
{
  "settings": {
    "analysis": {
      "analyzer": {
        "es_std":{
          "type":"standard",
          "stopwords":"_englist_"
        }
      }
    }
  }
}
           

修改后查看分词效果

//-----------用standard ,被分词成  a , dog , is , in , the , house
GET /my_index/_analyze
{
  "analyzer":"standard",
  "text":"a dog is in the house"
}


//----------用es_std, 被分词成   dog , house
GET /my_index/_analyze
{
  "analyzer": "es_std",
  "text": "a dog is in the house"
}
           

(3)定制化自己的分词器

PUT /my_index
{
  "settings": {
    "analysis": {
      "char_filter": {
        "&_to_and":{
          "type":"mapping",
          "mappings":["&=>and"]     //&转化成and
        }
      },
      "filter": {
        "my_stopwords":{
          "type":"stop",
          "stopwords":["the","a"]     //忽略the a
        }
      },
      "analyzer": {
        "my_analyzer":{
          "type":"custom",
          "char_filter":["html_strip","&_to_and"],    //忽略html标签,&转成and
          "tokenizer":"standard",
          "filter":["lowercase","my_stopwords"]
        }
      }
    }
  }
}



//--------------测试,分词结果:tomandjerry,are, friend,in, house,haha
GET /my_index/_analyze
{
  "analyzer": "my_analyzer",
  "text":"tom&jerry are a friend in the house,<a>,HAHA!!"
}
           

在type中应用自己的分词器

PUT /my_index/_mapping/my_type
{
  "properties": {
    "content":{
      "type": "text",
      "analyzer": "my_analyzer"
    }
  }
}
           

3、深入探秘type底层数据结构

(1)理论知识

  • type,是一个index中用来区分类似的数据的,类似的数据,但是可能有不同的fields,而且有不同的属性来控制索引建立、分词器
  • field的value,在底层的Lucene中建立索引的时候,全部是opaque(不透明)bytes类型,即:不区分类型的。
  • Lucene是没有type的概念的,在document中,实际上将type作为一个document的field来存储,即_type,es通过_type来进行type的过滤和筛选。
  • 一个index中的多个type,实际上是放在一起存储的,因此一个index下,不能有多个type重名,二类型或者其他设置不同的,因为那样是无法处理的。

(2)案例实践

(1)插入两条数据

PUT goods_index/electronic_goods/1
{
  "name": "geli kongtiao",
  "price": 1999.0,
  "service_period": "one year"
}

PUT goods_index/eat_goods/2
{
  "name": "aozhou dalongxia",
  "price": 199.0,
  "eat_period": "one week"
}

索引名称为goods_index,在该索引下面分别有两个type:electronic_goods和eat_goods

我们来看下索引对应的映射

(2)查看mapping

GET /goods_index/_mapping

{
  "goods_index": {
    "mappings": {
      "electronic_goods": {
        "properties": {
          "name": {
            "type": "text",
            "fields": {
              "keyword": {
                "type": "keyword",
                "ignore_above": 256
              }
            }
          },
          "price": {
            "type": "float"
          },
          "service_period": {
            "type": "text",
            "fields": {
              "keyword": {
                "type": "keyword",
                "ignore_above": 256
              }
            }
          }
        }
      },
      "eat_goods": {
        "properties": {
          "eat_period": {
            "type": "text",
            "fields": {
              "keyword": {
                "type": "keyword",
                "ignore_above": 256
              }
            }
          },
          "name": {
            "type": "text",
            "fields": {
              "keyword": {
                "type": "keyword",
                "ignore_above": 256
              }
            }
          },
          "price": {
            "type": "float"
          }
        }
      }
    }
  }
}

一个index中的多个type,实际上是放在一起存储的,在Lucene底层的存储结构如下:

(3)lucene 底层的存储

{
   "ecommerce": {
      "mappings": {
        "_type": {
          "type": "string",
          "index": "not_analyzed"
        },
        "name": {
          "type": "string"
        }
        "price": {
          "type": "double"
        }
        "service_period": {
          "type": "string"
        }
        "eat_period": {
          "type": "string"
        }
      }
   }
}

上述两条数据在底层存储结构如下:

{
  "_type": "electronic_goods",
  "name": "geli kongtiao",
  "price": 1999.0,
  "service_period": "one year",
  "eat_period": ""
}

{
  "_type": "eat_goods",
  "name": "aozhou dalongxia",
  "price": 199.0,
  "service_period": "",
  "eat_period": "one week"
}

_type字段就是type的名称,两个type中都有name字段,这里两个type中同有name字段,意味type同享一个存储空间,如果electronic_goods中的name为data类型,eat_goods中name为text类型。如果二者的类型不一样,这里就会存在问题,Lucene底层的数据结构会将“electronic_goods”和“eat_goods”的字段取并集存储起来,

           

(3)最后总结

将类似结构的type放在一个index下,这些type应该有多个field是相同的。假如说,你将两个type的field完全不同,放在一个index下,那么就每条数据都至少有一半的field在底层的Lucene中是空值,会有严重的性能问题。

4、mapping root object深入剖析

(1)root object:某个type对应的mapping json

就是某个type对应的mapping json,包括了properties,metadata(_id,_source,_type),settings(analyzer),其他settings(比如include_in_all)

(2)properties:包含type、index、analyzer

PUT  my_index/_mapping/my_type
{
        "properties":{
            "title":{
                "type":"string",     //类型
                "index":"analyzed",  //要不要分词
                "analyzer":"standard" //使用哪个分词器
            }
        }
}
           

(3)_source

好处:
  • 查询的时候,直接可以拿到完整的document,不需要先拿document id,再发送一次请求拿document
  • partial update基于_source实现
  • reindex时,直接基于_source实现,不需要从数据库(或者其他外部存储)查询数据再修改
  • 可以基于_source定制返回field
  • debug query更容易,因为可以直接看到_source

如果不需要上述好处,可以禁用_source:

禁用_source:

PUT /my_index/_mapping/my_type2
{
     "_source":{
          "enabled":false
     }

}
           

(4)_all

将所有field打包在一起,作为一个_all   field,建立索引。没指定任何field进行搜索时,就是使用_all field在搜索。

禁用 _all:

PUT  /my_index/_mapping/my_type3
{
        "_all":{
            "enabled":false
        }
}
           

也可以在field级别设置include_in_all field,设置是否要将field的值包含在_all field中

PUT  /my_index/_mapping/my_type4
{
    "properties":{
        "my_field":{
            "type":"text",
            "include_in_all":false
        }
    }
}
           

(5)标识性metadata:_index、_type、_id

5、定制化自己的dynamic mapping策略

(1)定制dynamic策略

  • true:遇到陌生字段,就进行dynamic mapping
  • false:遇到陌生字段,就忽略
  • strict:遇到陌生字段,就报错
//----------定制dynamic策略
PUT /my_index
{
  "mappings": {
    "my_type": {
      "dynamic":"strict",
      "properties": {
        "title":{
          "type": "text"
        },
        "address":{
          "type": "object",
          "dynamic":"true"
        }
      }
    }
  }
}

//---------1、测试顶层strict(遇到陌生字段报错),添加陌生字段content
PUT my_index/my_type/1
{
  "title":"my article",
  "content":"this is my article",
  "address":{
    "province":"guangdong",
    "city":"guangzhou"
  }
}

错误信息如下:
{
  "error": {
    "root_cause": [
      {
        "type": "strict_dynamic_mapping_exception",
        "reason": "mapping set to strict, dynamic introduction of [content] within [my_type] is not allowed"
      }
    ],
    "type": "strict_dynamic_mapping_exception",
    "reason": "mapping set to strict, dynamic introduction of [content] within [my_type] is not allowed"
  },
  "status": 400
}

//----------2、测试address(dynamic:true -- 遇到陌生字段进行dynamic mapping)中添加陌生字段
PUT my_index/my_type/1
{
  "title":"my article",
  "address":{
    "province":"guangdong",
    "city":"guangzhou"
  }
}

结果如下:
{
  "_index": "my_index",
  "_type": "my_type",
  "_id": "1",
  "_version": 1,
  "result": "created",
  "_shards": {
    "total": 2,
    "successful": 1,
    "failed": 0
  },
  "created": true
}


//-----------查看
GET /my_index/_mapping/my_type

结果如下:
{
  "my_index": {
    "mappings": {
      "my_type": {
        "dynamic": "strict",
        "properties": {
          "address": {
            "dynamic": "true",
            "properties": {
              "city": {
                "type": "text",
                "fields": {
                  "keyword": {
                    "type": "keyword",
                    "ignore_above": 256
                  }
                }
              },
              "province": {
                "type": "text",
                "fields": {
                  "keyword": {
                    "type": "keyword",
                    "ignore_above": 256
                  }
                }
              }
            }
          },
          "title": {
            "type": "text"
          }
        }
      }
    }
  }
}
           

(2)定制dynamic mapping策略

a、date_detection

默认会按照一定格式识别date,比如yyyy-MM-dd,但是如果某个field先过来一个2017-01-01的值,就会被自动dynamic mapping成date,后面如果再来一个hello world之类的值,就会报错,可以手动关闭某个type的date_detection,如果有需要,自己手动指定某个field为date类型

PUT  /my_index/_mapping/my_type

{

          "date_detection":false

}

b、定制自己的动态映射模板dynamic mapping template(type level)

//---------------1、定制模板:以“_en”结尾的字段,匹配如下mapping
PUT my_index
{
  "mappings": {
    "my_type": {
      "dynamic_templates":[
        {"en":{
          "match":"*_en",
          "match_mapping_type":"string",
          "mapping":{
            "type":"string",
            "analyzer":"english"
          }
        }}
        
        ]
    }
  }
}

//-----------------2、插入测试数据
PUT my_index/my_type/1
{
  "title":"this is my first article"
}

PUT my_index/my_type/2
{
  "title_en":"this is my first article"
}

//------------------3、查询
GET my_index/my_type/_search
{
  "query": {
    "match": {
      "title": "is"
    }
  }
}

结果:
{
  "took": 145,
  "timed_out": false,
  "_shards": {
    "total": 5,
    "successful": 5,
    "failed": 0
  },
  "hits": {
    "total": 1,
    "max_score": 0.2824934,
    "hits": [
      {
        "_index": "my_index",
        "_type": "my_type",
        "_id": "1",
        "_score": 0.2824934,
        "_source": {
          "title": "this is my first article"
        }
      }
    ]
  }
}


GET my_index/my_type/_search
{
  "query": {
    "match": {
      "title_en": "is"
    }
  }
}

结果:
{
  "took": 3,
  "timed_out": false,
  "_shards": {
    "total": 5,
    "successful": 5,
    "failed": 0
  },
  "hits": {
    "total": 0,
    "max_score": null,
    "hits": []
  }
}
           

title没有匹配到任何的dynamic模板,默认就是standard分词器,不会过滤停用词,is会进入倒排索引,用is来搜索是可以搜索到的;title_en匹配到了dynamic模板,就是English分词器,会过滤停用词,is这种停用词就会被过滤掉,用is来搜索就搜索不到了

c、定制自己的default mapping template(index level)

PUT /my_index
{
    "mappings":{
        "_default":{
          "_all":{"enabled":false}
        },
        "blog":{
          "_all":{"enabled": false}
        }
    }
}
           

6、基于scroll滚动搜索和alias别名实现zero downtime reindex

(1)重建索引

一个field的设置是不能被修改的,如果要修改一个field,那么应该重新按照新的mapping,建立一个index,然后将数据批量查询出来,重新用bulk api写入index中,批量查询的时候,建议采用scroll api,并且采用多线程并发的方式来reindex数据,每次scroll就查询指定日期的一段数据,交给一个线程即可。

(1)一开始,依靠dynamic mapping,插入数据,但是不小心有些数据是2017-01-01这种日期格式的,所以title这种field被自动映射为了date类型,实际上他应该是string类型
PUT my_index/my_type/1
{
  "title":"2017-01-01"
}
PUT my_index/my_type/2
{
  "title":"2017-01-02"
}
PUT my_index/my_type/3
{
  "title":"2017-01-03"
}

GET my_index/_mapping/my_type
结果:
{
  "my_index": {
    "mappings": {
      "my_type": {
        "properties": {
          "title": {
            "type": "date"
          }
        }
      }
    }
  }
}

(2)当后期向索引中加入string类型的title值时,就会报错
PUT my_index/my_type/4
{
  "title":"my first article"
}

结果:
{
  "error": {
    "root_cause": [
      {
        "type": "mapper_parsing_exception",
        "reason": "failed to parse [title]"
      }
    ],
    "type": "mapper_parsing_exception",
    "reason": "failed to parse [title]",
    "caused_by": {
      "type": "illegal_argument_exception",
      "reason": "Invalid format: \"my first article\""
    }
  },
  "status": 400
}

(3)如果此时想修改title的类型,是不可能的
PUT my_index/_mapping/my_type
{
  "properties": {
    "title":{
      "type": "text"
    }
  }
}

结果:
{
  "error": {
    "root_cause": [
      {
        "type": "illegal_argument_exception",
        "reason": "mapper [title] of different type, current_type [date], merged_type [text]"
      }
    ],
    "type": "illegal_argument_exception",
    "reason": "mapper [title] of different type, current_type [date], merged_type [text]"
  },
  "status": 400
}

(4)此时,唯一的办法就是进行reindex,也就是说,重新建立一个索引,将旧索引的数据查询出来,在导入新索引
(5)如果说旧索引的名字,是old_index,新索引的名字是new_index,终端java应用,已经在使用old_index在操作了,难道还要去停止java应用,修改使用的index为new_index,才重新启动java应用吗?这个过程中,就会导致java应用停机,可用性降低
(6)所以说,给java应用一个别名,这个别名是指向旧索引的,java应用先用着,java应用先用goods_index alias来操作,此时实际指向的是旧的my_index

PUT my_index/_alias/goods_index

(7)新建一个index,调整其title的类型为string
PUT my_index_new
{
  "mappings": {
    "my_type":{
      "properties": {
        "title":{
          "type": "text"
        }
      }
    }
  }
}

(8)使用scroll api将数据批量查询出来
GET my_index/_search?scroll=1m
{
  "query": {
    "match_all": {}
  },
  "sort": ["_doc"],
  "size": 1
}

结果:
{
  "_scroll_id": "DnF1ZXJ5VGhlbkZldGNoBQAAAAAAABvLFjU4LWIxRl9CU1BLR1RHMHd6SWpuX3cAAAAAAAAbyRY1OC1iMUZfQlNQS0dURzB3eklqbl93AAAAAAAAG8oWNTgtYjFGX0JTUEtHVEcwd3pJam5fdwAAAAAAABvMFjU4LWIxRl9CU1BLR1RHMHd6SWpuX3cAAAAAAAAbzRY1OC1iMUZfQlNQS0dURzB3eklqbl93",
  "took": 113,
  "timed_out": false,
  "_shards": {
    "total": 5,
    "successful": 5,
    "failed": 0
  },
  "hits": {
    "total": 3,
    "max_score": null,
    "hits": [
      {
        "_index": "my_index",
        "_type": "my_type",
        "_id": "2",
        "_score": null,
        "_source": {
          "title": "2017-01-02"
        },
        "sort": [
          0
        ]
      }
    ]
  }
}

(9)采用bulk api将scroll查出来的一批数据,批量写入新索引
POST /_bulk
{"index":{"_index":"my_index_new", "_type":"my_type","_id":"2"}}
{"title":"2017-01-02"}

(10)反复循环(8)-(9),查询一批又一批的数据出来,采取bulk api将每一批数据批量写入新索引
(11)将goods_index alias切换到my_index_new上去,java应用会直接通过index别名使用新的索引中的数据,java应用程序不需要停机,零停机,高可用

POST /_aliases
{
  "actions": [
    {"remove": {"index": "my_index","alias": "goods_index"}},
    {"add": {"index": "my_index_new","alias": "goods_index"}}
  ]
}

(12)直接通过goods_index别名来查询

GET /goods_index/my_type/_search
           

继续阅读