ElsticSearch常用指南

群日常问题
动态字段的问题nested?
经测试不支持list<map> 和 list<? extend map>
支持指定字段在es的索引中忽略大小写,以便在term查询时不区分大小写,仅对keyword类型字段生效,es的规则,并非框架限制.
Normalizers是Elasticsearch中用于文本规范化的一种组件，它们通常用于关键字字段（keyword fields），这些字段需要规范化但不需要分词。Normalizers和analyzers类似，但只能包含字符过滤器（char_filter）和标记过滤器（token filter），而不包含分词器（tokenizer）。
PUT test0612
# 关闭后只能读
POST test0612/_close
PUT /test0612/_settings
{
  "index":{
    "analysis":{
      "normalizer":{
        "lowercase_normalizer":{
          "type": "custom",
          "char_filter": [],
           "filter": ["lowercase"]
        }
      }
    }
  }
}
POST test0612/_open

PUT /test0612/_mapping
{
  "properties": {
    "name": {
      "type": "keyword",
      "normalizer": "lowercase_normalizer"
    },
    "preName": {
      "type": "keyword"
    }
  }
}

POST _analyze
{
  "text": "TEST测试",
  "analyzer": "ik_max_word"
}

GET test0612/_analyze
{
  "field": "name",
  "text": "TEST测试"
}

GET test0612/_analyze
{
  "field": "preName",
  "text": "TEST测试"
}
测试

PUT test0612/_doc/1
{
  "id": "1",
  "name": "Mohammed Ahmed",
  "preName": "Mohammed Ahmed"
}
PUT test0612/_doc/2
{
"id": "2",
"name": "Sophie Martin",
"preName": "Mohammed Ahmed"
}
PUT test0612/_doc/3
{
  "id": "3",
  "name": "mohammed ahmed",
  "preName": "Mohammed Ahmed"
}

GET test0612/_search
{
  "query": {
    "wildcard": {
      "name": {
        "value": "*moham*"
      }
    }
  }
}

GET test0612/_search
{
  "query": {
    "term": {
      "name": {
        "value": "MohAmmed Ahmed"
      }
    }
  }
}

join属于文档间建立父子关系，若一个索引中通过某个字段建立的父子关系，该如何使用呢？
Nested
首先要说nested,为什么要nested, 如果不指定nested, 那么默认的类型是Object, 其內部结构是平铺类型，无法排序和聚合，
默认情况下的平铺查询方式
# 平铺查询会导致查询到非张四的数据
GET /test_0611/_search
{
  "query": {
    "bool": {
       "must": [
        {
          "match": {
            "users.name": "张"
          }
        },
        {
          "match": {
            "users.name": "四"
          }
        }
      ]
    }
  }
}
嵌套
POST /test_0611/_mapping
{
  "properties": {
    "users": {
      "type": "nested"
    }
  }
}
POST /test_0611/_doc
{
     "users": [
          {
               "name": "张 三",
               "age": 23,
               "email": "5@qq.com"
          },
          {
               "name": "李 四",
               "age": 18,
               "email": "6@qq.com"
          }
     ]
}
# 查询不到数据
GET /test_0611/_search
{
  "query": {
    "nested": {
      "path": "users",
      "query": {
        "bool": {
          "must": [
            {
              "match": {
                "users.name": {
                  "query": "张",
                  "operator": "and"
                }
              }
            },
            {
              "match": {
                "users.name": {
                  "query": "四",
                  "operator": "and"
                }
              }
            }
          ]
        }
      }
    }
  }
}
JOIN
DELETE my_blogs

PUT my_blogs
{
    "settings": {
        "number_of_shards": 2
    },
    "mappings": {
        "properties": {
            "blog_comment_relation": {
                "type": "join",
                "relations": {
                    "blog": "comment"
                }
            },
            "content": {
                "type": "text",
                "analyzer": "ik_max_word",
                "search_analyzer": "ik_smart"
            },
            "title": {
                "type": "keyword"
            }
        }
    }
}

#! name 标记为父文档, 指定id = blog1
PUT my_blogs/_doc/blog1
{
  "title": "中间人还有未来吗?",
  "content": "中间人还有未来吗?我觉的应该有",
  "blog_comment_relation": {
    "name": "blog"
  }
}

#! 由es生成Id; YiBtCpAB3lg0W2IdEzLJ
POST my_blogs/_doc
{
  "title": "宇宙的尽头是什么?",
  "content": "或许是灵魂在银河里畅游",
  "blog_comment_relation": {
    "name": "blog"
  }
}

#! 指定id=comment1
PUT my_blogs/_doc/comment1?routing=blog1
{
  "title": "精彩",
  "content": "你的文章很有意思",
  "user":"王二小",
  "blog_comment_relation": {
    "name": "comment",
    "parent": "blog1"
  }
}
# 不指定Id
POST my_blogs/_doc?routing=YiBtCpAB3lg0W2IdEzLJ
{
  "title": "不同意",
  "content": "宇宙的尽头应该还是宇宙",
  "user":"张三",
  "blog_comment_relation": {
    "name": "comment",
    "parent": "YiBtCpAB3lg0W2IdEzLJ"
  }
}
设计父子结构时要考虑将父子的所有字段能考虑进去
# 父和子的所有字段
GET /my_blogs/_mapping
# 父和子的所有文档
GET /my_blogs/_search 
{
  "query": {"match_all": {}}
}
# 父文档查询， 不会包含子文档
GET my_blogs/_doc/blog1

# 通过parent_id查询子文档
POST my_blogs/_search
{
  "query": {
    "parent_id": {
      "type": "comment",
      "id": "blog1"
    }
  }
}

# 使用has_child，通过child查询parent
# 要求有孩子类型comment , 并且孩子 content match 文章
POST my_blogs/_search
{
  "query": {
    "has_child": {
      "type": "comment",
      "query": {
        "match": {
          "content": "文章"
        }
      }
    }
  }
}

# 使用has_parent，查询关联的子文档
POST my_blogs/_search
{
  "query": {
    "has_parent": {
      "parent_type": "blog",
      "query": {
        "match": {
          "content": "中间"
        }
      }
    }
  }
}
# 不允许根据子ID查询
GET my_blogs/_doc/comment1
这样是可以的
GET my_blogs/_doc/comment1?routing=blog1