概要

elasticsearch一直在使用，这里总结一下mappings的修改方法，分为两种情况：

1. 增加新的字段，这种很简单；
2. 修改已有的字段类型，这种就比较麻烦了，需要reindex,对索引进行迁移重建。

一、创建索引

curl -XPUT 'http://127.0.0.1:9200/test?pretty' -H  'Content-Type: application/json' -d '{"settings":{},"mappings":{}}'

{
	"settings": {
		"index": {
			"number_of_shards": 2,
			"number_of_replicas": 3,
			"analysis": {
				"analyzer": {
					"char_analyzer": {
						"filter": ["lowercase"],
						"type": "custom",
						"tokenizer": "char_split"
					}
				},
				"tokenizer": {
					"char_split": {
						"token_chars": ["letter", "digit", "punctuation", "symbol"],
						"min_gram": "1",
						"type": "nGram",
						"max_gram": "1"
					}
				}
			}
		}
	},
	"mappings": {
		"doc": {
			"properties": {
				"id": {
					"type": "long"
				},
				"pd_name": {
					"type": "text",
					"fields": {
						"keyword": {
							"type": "keyword",
							"ignore_above": 256
						}
					},
					"analyzer": "char_analyzer"
				},
				"pd_uname": {
					"type": "text",
					"fields": {
						"keyword": {
							"type": "keyword",
							"ignore_above": 256
						}
					}
				},
				"product_id": {
					"type": "long"
				}
			}
		}
	}
}

1.1、获取mappings

curl -XGET 'http://127.0.0.1:9200/test/_mappings?pretty'
{
  "test" : {
    "mappings" : {
      "doc" : {
        "properties" : {
          "id" : {
            "type" : "long"
          },
          "pd_name" : {
            "type" : "text",
            "fields" : {
              "keyword" : {
                "type" : "keyword",
                "ignore_above" : 256
              }
            },
            "analyzer" : "char_analyzer"
          },
          "pd_uname" : {
            "type" : "text",
            "fields" : {
              "keyword" : {
                "type" : "keyword",
                "ignore_above" : 256
              }
            }
          },
          "product_id" : {
            "type" : "long"
          }
        }
      }
    }
  }
}

二、新增字段修改mappings

增加一个 new_stocks 字段，如下：

curl -XPUT 'http://127.0.0.1:9200/test/doc/_mapping?pretty' -H  'Content-Type: application/json' -d '{"properties":{"new_stocks":{"type":"nested","properties":{"value":{"type":"long"},"conversion":{"type":"long"}}}}}'
{
  "acknowledged" : true
}

再查一下：

curl -XGET 'http://127.0.0.1:9200/test/_mappings?pretty'                                                                                                          {
  "test" : {
    "mappings" : {
      "doc" : {
        "properties" : {
          "id" : {
            "type" : "long"
          },
          "new_stocks" : {
            "type" : "nested",
            "properties" : {
              "conversion" : {
                "type" : "long"
              },
              "value" : {
                "type" : "long"
              }
            }
          },
          "pd_name" : {
            "type" : "text",
            "fields" : {
              "keyword" : {
                "type" : "keyword",
                "ignore_above" : 256
              }
            },
            "analyzer" : "char_analyzer"
          },
          "pd_uname" : {
            "type" : "text",
            "fields" : {
              "keyword" : {
                "type" : "keyword",
                "ignore_above" : 256
              }
            }
          },
          "product_id" : {
            "type" : "long"
          }
        }
      }
    }
  }
}

可以看到new_stocks字段已经加上去了。

三、修改OR删除mappings已有字段

如果想把product_id字段类型由long改成text，并删除id字段呢？此时用第二章节的方案就不行了，需要借用reindex命令重做索引。

3.1、创建新索引，将要改字段加进去

curl -XPUT 'http://127.0.0.1:9200/new_test?pretty' -H  'Content-Type: application/json' -d '{"settings":{},"mappings":{}}'

{
	"settings": {
		"index": {
			"number_of_shards": 2,
			"number_of_replicas": 3,
			"analysis": {
				"analyzer": {
					"char_analyzer": {
						"filter": ["lowercase"],
						"type": "custom",
						"tokenizer": "char_split"
					}
				},
				"tokenizer": {
					"char_split": {
						"token_chars": ["letter", "digit", "punctuation", "symbol"],
						"min_gram": "1",
						"type": "nGram",
						"max_gram": "1"
					}
				}
			}
		}
	},
	"mappings": {
		"doc": {
			"properties": {
				"id": {
					"type": "long"
				},
				"pd_name": {
					"type": "text",
					"fields": {
						"keyword": {
							"type": "keyword",
							"ignore_above": 256
						}
					},
					"analyzer": "char_analyzer"
				},
				"pd_uname": {
					"type": "text",
					"fields": {
						"keyword": {
							"type": "keyword",
							"ignore_above": 256
						}
					}
				},
				"product_id": {
					"type": "text",
					"fields": {
						"keyword": {
							"type": "keyword",
							"ignore_above": 256
						}
					}
				}
			}
		}
	}
}

3.2、同步数据

curl -XPOST 'http://127.0.0.1:9200/_reindex?pretty' -H  'Content-Type: application/json' -d '{"source":{"index":"test"},"dest":{"index":"new_test"}}'

3.3、删除原索引并对新索引重命名

#删除
curl -XDELETE 'http://127.0.0.1:9200/test?pretty'
#设置别名
curl -XPOST 'http://127.0.0.1:9200/_aliases?pretty' -H  'Content-Type: application/json' -d '{"actions":[{"add":{"index":"new_test","alias":"test"}}]}'

3.4、同步数据的技巧

1. 同步部分字段

#从源头过滤, _source 参数，只同步 id，pd_name两个字段
curl -XPOST 'http://127.0.0.1:9200/_reindex?pretty' -H  'Content-Type: application/json' -d '{"source":{"index":"test","_source":["id","pd_name"]},"dest":{"index":"new_test"}}'
#从目的地移除 脚本控制，移除id字段
curl -XPOST 'http://127.0.0.1:9200/_reindex?pretty' -H  'Content-Type: application/json' -d '{"source":{"index":"test"},"dest":{"index":"new_test"},"script":{"lang":"groovy","inline":"ctx._source.remove(\"id\");ctx._source.remove(\"pd_uname\")"}}'

2. 提高同步速率
   调整参数size大小，默认1000每批

curl -XPOST 'http://127.0.0.1:9200/_reindex?pretty' -H  'Content-Type: application/json' -d '{"source":{"index":"test","size":5000},"dest":{"index":"new_test"}}'

四、参考文献

1]:官方文档
2]:ES索引重建reindex详解
3]:ES迁移效率