参考：https://blog.csdn.net/qq_26230421/article/details/80366649

集群

  1. # 查看集群是否健康
  2. GET _cat/health?v
  4. # 查看集群节点
  5. GET _cat/nodes?v
  7. # 查看有什么索引
  8. GET _cat/indices

索引

一个分片就是一个Lucene索引，根据自己情况合理设置，每30G（20亿数据）新增一个分片。 备份数可以提高查询性能，根据节点数合理设置，一般1-3个。

  • 创建
  1. PUT /people
  2. {
  3.   "settings": {
  4.     "number_of_shards": 1, # 分片数
  5.     "number_of_replicas": 0, # 备份数
  6.         "index.analysis.analyzer.default.type": "ik_max_word" # 自定义默认分词器，默认：english
  7.   }
  8. }
  • 删除
  1. DELETE /people
  3. # 也可以删除多个索引
  4. DELETE /index_one,index_two

mapping

  1. 追加或设置mapping
  2. PUT /people/_mapping/doc
  3. {
  4.     "dynamic": "strict",
  5.   "properties": {
  6.     "name": {
  7.       "type": "text",
  8.       "analyzer": "ik_max_word", # 索引分词器，如果不加，则text会使用默认分词器,
  9.       "search_analyzer": "ik-smart" # 搜索分词器，如果不加，默认和索引分词器一致。
  10.     },
  11.     "age": {
  12.       "type": "integer"
  13.     },
  14.     "birthday": {
  15.       "type": "date",
  16.       "format": "yyyy-MM-dd HH:mm:ss||yyyy-MM-dd||epoch_millis"
  17.     }
  18.   }
  19. }
  21. # 可以为字段设置keyword（方便精确匹配）
  22. PUT /people/_mapping/man
  23. {
  24.   "properties": {
  25.     "question":{
  26.       "type": "text",
  27.       "fields": {
  28.         "keyword":{
  29.           "type": "keyword"
  30.         }
  31.       }
  32.     }
  33.   }
  34. }
  37. # 查看mapping
  38. GET /people/_mapping

“dynamic”: “strict” 表示如果遇到陌生field会报错 “dynamic”: “true” 表示如果遇到陌生字段，就进行dynamic mapping

“dynamic”: “false” 表示如果遇到陌生字段，就忽略

别名

  1. # 创建别名
  2. PUT /people/_alias/test
  4. # 查询索引的别名
  5. GET /people/_alias
  6. >>>
  7. {
  8.   "people": {
  9.     "aliases": {
  10.       "test": {}
  11.     }
  12.   }
  13. }
  14. # 查询别名指向哪一个索引
  15. GET /*/_alias/test
  16. >>>
  17. {
  18.   "people": {
  19.     "aliases": {
  20.       "test": {}
  21.     }
  22.   }
  23. }

文档

增加

put必须设置id，post则不用

  1. PUT /people/man/1
  2. {
  3.   "name": "叶良辰",
  4.   "country": "china",
  5.   "age": 25,
  6.   "birthday": "1993-01-01",
  7.   "desc": "叶良辰，本地人，狂妄自大，惹了他，会有一百种方法让你呆不下去！与赵日天是好基友，两个人风风火火闯九州"
  8. }
  9. 》》》
  10. {
  11.   "_index": "people",
  12.   "_type": "man",
  13.   "_id": "1",
  14.   "_version": 1,
  15.   "result": "created",
  16.   "_shards": {
  17.     "total": 1,
  18.     "successful": 1,
  19.     "failed": 0
  20.   },
  21.   "created": true
  22. }
  1. POST /people/man
  2. {
  3.   "name": "叶良辰",
  4.   "country": "china",
  5.   "age": 25,
  6.   "birthday": "1993-01-01",
  7.   "desc": "叶良辰，本地人，狂妄自大，惹了他，会有一百种方法让你待不下去！与赵日天是好基友，两个人风风火火闯九州"
  8. }
  9. 》》》
  10. {
  11.   "_index": "people",
  12.   "_type": "man",
  13.   "_id": "AW7P2HV23ydeaETTj2Cz",
  14.   "_version": 1,
  15.   "result": "created",
  16.   "_shards": {
  17.     "total": 1,
  18.     "successful": 1,
  19.     "failed": 0
  20.   },
  21.   "created": true
  22. }

修改

  • 全文修改使用的是PUT命令，把所有字段都带上
  1. PUT /people/man/1
  2. {
  3.   "name": "叶良辰",
  4.   "country": "china",
  5.   "age": 26,
  6.   "birthday": "1993-01-01",
  7.   "desc": "叶良辰，本地人，狂妄自大，惹了他，会有一百种方法让你呆不下去！与赵日天是好基友，两个人风风火火闯九州"
  8. }
  9. 》》》
  10. {
  11.   "_index": "people",
  12.   "_type": "man",
  13.   "_id": "1",
  14.   "_version": 2,
  15.   "result": "updated",
  16.   "_shards": {
  17.     "total": 1,
  18.     "successful": 1,
  19.     "failed": 0
  20.   },
  21.   "created": false
  22. }
  • 部分修改使用POST
  1. POST /people/man/1/_update
  2. {
  3.   "doc": {
  4.     "name": "zhangsan"
  5.   }
  6. }
  7. 》》》
  8. {
  9.   "_index": "people",
  10.   "_type": "man",
  11.   "_id": "1",
  12.   "_version": 3,
  13.   "result": "updated",
  14.   "_shards": {
  15.     "total": 1,
  16.     "successful": 1,
  17.     "failed": 0
  18.   }
  19. }
  • 脚本修改
  1. POST /people/man/1/_update
  2. {
  3.   "script": "ctx._source.age += 10"
  4. }

删除

  1. DELETE /people/man/1
  3. POST /people/man/_delete_by_query?conflicts=proceed
  4. {
  5.   "query": {
  6.     "match_all": {}
  7.   }
  8. }

查询

简单查询

match分词查询

  1. 方式一（简单查询）：
  2. GET /people/_search
  4. 方式二：
  5. POST /people/_search
  6. {
  7.   "query": {
  8.     "match_all": {}
  9.   }
  10. }
  12. GET people/man/_search?_source=name,country
  13. {
  14.   "query": {
  15.     "match": {
  16.       "age": "25"
  17.     }
  18.   }
  19. }

排序

  1. POST /people/_search
  2. {
  3.   "query": {
  4.     "match": {
  5.       "name": "叶良辰"
  6.     }
  7.   },
  8.   "sort": [
  9.     {
  10.       "birthday": {
  11.         "order": "desc"
  12.       }
  13.     }
  14.   ]
  15. }

多字段搜索

  1. GET /people/_search
  2. {
  3.   "query": {
  4.     "multi_match": {
  5.       "query": "页良",
  6.       "fields": ["name","desc"]
  7.     }
  8.   }
  9. }

match_phrase：分词，并且要全部包含

  1. 完全匹配可能比较严，我们会希望有个可调节因子，少匹配一个也满足，那就需要使用到slop。
  2. {
  3.   "query": {
  4.     "match_phrase": {
  5.         "content" : {
  6.             "query" : "我的宝马多少马力",
  7.             "slop" : 1
  8.         }
  9.     }
  10.   }
  11. }

query_string

  1. GET /people/_search
  2. {
  3.   "query": {
  4.     "query_string": {
  5.       "default_field": "desc",
  6.       "query": "(叶良辰 AND 风)"
  7.     }
  8.   }
  9. }
  11. “叶良辰 AND 风”会先把“叶良辰”拆分成“叶”、“良”和“辰”，然后后面必须有“风”
  13. GET /people/_search
  14. {
  15.   "query": {
  16.     "query_string": {
  17.       "query": "(叶良辰 AND 火) OR (赵日天 AND 风)",
  18.       "fields": ["name","desc"]
  19.     }
  20.   }
  21. }

term 

  1. GET /people/_search
  2. {
  3.   "query": {
  4.     "term": {
  5.       "name": "叶良辰"
  6.     }
  7.   }
  8. }

分页

  1. GET people/_search
  2. {
  3.   "query": {
  4.     "match_all": {}
  5.   },
  6.   "from": 0,
  7.   "size": 1
  8. }

范围查询

数字

  1. GET /people/_search
  2. {
  3.   "query": {
  4.     "range": {
  5.       "age": {
  6.         "gte": 10,
  7.         "lte": 26
  8.       }
  9.     }
  10.   }
  11. }

日期

  1. GET /people/_search
  2. {
  3.   "query": {
  4.     "range": {
  5.       "birthday": {
  6.         "gte": "1993-01-01",
  7.         "lte": "now"
  8.       }
  9.     }
  10.   }
  11. }

过滤查询

过滤很快，但是没有评分，后面的boost就是默认的分数。

  1. POST /people/man/_search
  2. {
  3.   "query": {
  4.     "constant_score": {
  5.       "filter": {
  6.         "range": {
  7.           "age": {
  8.             "gte": 20,
  9.             "lte": 30
  10.           }
  11.         }
  12.       },
  13.       "boost": 1.2
  14.     }
  15.   }
  16. }

也可以使用bool must 过滤

  1. GET /city/_search
  2. {
  3.   "query": {
  4.     "bool": {
  5.       "must": [
  6.         {"match_all": {}}
  7.       ], 
  8.       "filter": {
  9.         "geo_distance": {
  10.           "distance": "1000km",
  11.           "location": {
  12.             "lat": 23.73,
  13.             "lon": 112.1
  14.           }
  15.         }
  16.       }
  17.     }
  18.   }
  19. }

布尔查询

shoud 相当于 or
must 相当于and

  1. POST /people/_search
  2. {
  3.   "query": {
  4.     "bool": {
  5.       "should": [
  6.         {
  7.           "match": {
  8.             "name": "叶良辰"
  9.           }
  10.         },
  11.         {
  12.           "match": {
  13.             "desc": "赵日天"
  14.           }
  15.         }
  16.       ]
  17.     }
  18.   }
  19. }
  1. POST /people/_search
  2. {
  3.   "query": {
  4.     "bool": {
  5.       "must": [
  6.         {
  7.           "match": {
  8.             "name": "叶良辰"
  9.           }
  10.         },
  11.         {
  12.           "match": {
  13.             "desc": "赵日天"
  14.           }
  15.         }
  16.       ]
  17.     }
  18.   }
  19. }

高亮

  1. GET /people/_search
  2. {
  3.   "query": {
  4.     "match": {
  5.       "name": "叶良辰2"
  6.     }
  7.   },
  8.   "highlight": {
  9.     "fields": {
  10.       "name": {}
  11.     }
  12.   }
  13. }

效果：

image.png

当前，也可以指定标签（虽然感觉没啥用）

  1. GET people/_search
  2. {
  3.   "query": {
  4.     "match": {
  5.       "name": "叶良辰"
  6.     }
  7.   },
  8.   "highlight": {
  9.     "pre_tags": ["<b>"], 
  10.     "post_tags": ["</b>"], 
  11.     "fields": {
  12.       "name": {}
  13.     }
  14.   }
  15. }

聚合

根据字段类型查询

  1. GET /people/_search
  2. {
  3.   "query": {
  4.     "match": {
  5.       "name": "叶良辰"
  6.     }
  7.   },
  8.   "aggs": {
  9.     "groupByAge": {
  10.       "terms": {
  11.         "field": "age",
  12.         "size": 10
  13.       }
  14.     }
  15.   },
  16.   "size": 0
  17. }
  19. >>>
  21. {
  22. …………………………………………………………………………
  23.   "aggregations": {
  24.     "groupByAge": {
  25.       "doc_count_error_upper_bound": 0,
  26.       "sum_other_doc_count": 0,
  27.       "buckets": [
  28.         {
  29.           "key": 25,
  30.           "doc_count": 5
  31.         },
  32.         {
  33.           "key": 26,
  34.           "doc_count": 1
  35.         }
  36.       ]
  37.     }
  38.   }
  39. }

统计：求和、平均值

  1. GET /people/_search
  2. {
  3.   "query": {
  4.     "match": {
  5.       "name": "叶良辰"
  6.     }
  7.   },
  8.   "aggs": {
  9.     "groupByAge": {
  10.       "stats": {    // 可以替换为min等
  11.         "field": "age"
  12.       }
  13.     }
  14.   },
  15.   "size": 0
  16. }
  18. 》》》
  20. "aggregations": {
  21.   "groupByAge": {
  22.     "count": 6,
  23.     "min": 25,
  24.     "max": 26,
  25.     "avg": 25.166666666666668,
  26.     "sum": 151
  27.   }
  28. }

先分组再计算

  1. GET /people/_search
  2. {
  3.   "aggs": {
  4.     "groupByAge": {
  5.       "terms": {
  6.         "field": "key"
  7.       },
  8.       "aggs": {
  9.         "avg-age": {
  10.           "avg": {
  11.             "field": "age"
  12.           }
  13.         }
  14.       }
  15.     }
  16.   },
  17.   "size": 0
  18. }
  20. 》》》
  21. "aggregations": {
  22.   "groupByAge": {
  23.     "doc_count_error_upper_bound": 0,
  24.     "sum_other_doc_count": 0,
  25.     "buckets": [
  26.       {
  27.         "key": "叶良辰",
  28.         "doc_count": 1,
  29.         "avg-age": {
  30.           "value": 26
  31.         }
  32.       }
  33.     ]
  34.   }
  35. }
  38. 提示：
  39. group字段必须没有分词，并且设置fielddata,如下，否则会报错(不一定）
  41. POST /people/_mapping/man
  42. {
  43.   "properties": {
  44.     "name":{
  45.       "type": "text",
  46.       "fielddata": true
  47.     }
  48.   }
  49. }

深度排序

实时查询：使用search-after

第一步正常查询，但是排序的时候，多加一个字段_uid：

  1. POST /people/_search
  2. {
  3.   "query": {
  4.     "match": {
  5.       "name": "叶良辰"
  6.     }
  7.   },
  8.   "sort": [
  9.     {
  10.       "birthday": {
  11.         "order": "desc"
  12.       },
  13.       "_uid":{
  14.         "order": "desc"
  15.       }
  16.     }
  17.   ]
  18. }
  20. 》》》
  21.   "_index": "people",
  22.   "_type": "man",
  23.   "_id": "AW7P7ag_3ydeaETTj2C3",
  24.   "_score": null,
  25.   "_source": {
  26.     "name": "叶良辰",
  27.     "country": "china",
  28.     "age": 25,
  29.     "birthday": "1993-01-01",
  30.     "desc": "叶良辰，本地人，狂妄自大，惹了他，会有一百种方法让你待不下
  31. 日天是好基友，两个人风风火火闯九州"
  32.   },
  33.   "sort": [
  34.     725846400000,
  35.     "man#AW7P7ag_3ydeaETTj2C3"
  36.   ]
  37. },

返回的结果里面会sort的值，

第二步查询的时候加上第一次返回的最后的值，然后就OK啦

  1. POST /people/_search
  2. {
  3.   "query": {
  4.     "match": {
  5.       "name": "叶良辰"
  6.     }
  7.   },
  8.   "sort": [
  9.     {
  10.       "birthday": {
  11.         "order": "desc"
  12.       },
  13.       "_uid":{
  14.         "order": "desc"
  15.       }
  16.     }
  17.   ],
  18.   "search_after": [725846400000, "man#AW7QFn6r3ydeaETTj2C8"]
  19. }

后台任务可以使用scroll
**
相当于获取一个镜像，然后可以随意处理

第一步

  1. POST /people/_search?scroll=1m
  2. {
  3.   "query": {
  4.     "match": {
  5.       "name": "叶良辰"
  6.     }
  7.   },
  8.   "sort": "_doc",
  9.   "size": 2
  10. }

获得scroll_id

image.png

第二步或者第N步（id不会变化）

  1. POST /_search/scroll
  2. {
  3.   "scroll": "1m",
  4.   "scroll_id":"DnF1ZXJ5VGhlbkZldGNoAwAAAAAACoDvFmpIaWxISVUxUUVDZUVpS1E4VXNaWmcAAAAAAAqA7hZqSGlsSElVMVFFQ2VFaUtROFVzWlpnAAAAAAAKgPAWakhpbEhJVTFRRUNlRWlLUThVc1paZw=="
  5. }

分词查询

  1. GET  _analyze
  2. {
  3.   "analyzer": "ik_smart",
  4.   "text": ["今天天气不错"]
  5. }
  7. GET  _analyze
  8. {
  9.   "analyzer": "standard",
  10.   "text": ["今天天气不错"]
  11. }

地理位置查询

设置索引

  1. PUT /city
  2. {
  3.   "settings": {
  4.     "number_of_shards": 1,
  5.     "number_of_replicas": 0
  6.   },
  7.   "mappings": {
  8.     "doc":{
  9.       "properties":{
  10.         "city":{
  11.           "type":"text"
  12.         },
  13.         "state": {"type": "text"},
  14.         "location":{"type":"geo_point"}
  15.       }
  16.     }
  17.   }
  18. }

插入数据

  1. POST /city/doc
  2. {
  3.   "city": "Guangzhou",
  4.   "state": "GD",
  5.   "location": {
  6.     "lat": "23.16667",
  7.     "lon": "113.23333"
  8.   }
  9. }

查询

  1. GET /city/_search
  2. {
  3.   "query": {
  4.     "bool": {
  5.       "must": [
  6.         {"match_all": {}}
  7.       ], 
  8.       "filter": {
  9.         "geo_distance": {
  10.           "distance": "1000km",
  11.           "location": {
  12.             "lat": 23.73,
  13.             "lon": 112.1
  14.           }
  15.         }
  16.       }
  17.     }
  18.   }
  19. }