1. 创建新表
curl -XPUT 'host:port/table_name?pretty'
table_name:表名,不能有大写字母,不能以下划线开头
pretty:使返回结果以便于阅读的JSON格式返回
自定义 schema,即mapping两种方式:
1.在创建表的时候一块创建:
host:port/table_name
{
"settings": {
"number_of_shards":4,
"number_of_replicas":1
},
"mappings": {
"dishlist": {
"_all": {
"enabled":false
},
"properties": {
"id": {
"type":"string",
"index":"not_analyzed"
},
"dish_name": {
"type":"string",
"index":"not_analyzed"
},
"dish_name_analyzed": {
"type":"string",
"index":"analyzed",
"analyzer":"ik_max_word"
},
"dish_id": {
"type":"string",
"index":"not_analyzed"
},
"poi_id": {
"type":"long",
"index":"not_analyzed"
},
"poi_name": {
"type":"string",
"store":"true"
}
}
}
}
}
2.在创建type的时候创建
host:port/table_name/type_name/_mapping?pretty
{
"type_name": {
"_all": {
"enabled":false
},
"properties": {
"id": {
"type":"string",
"index":"not_analyzed"
},
"dish_name": {
"type":"string",
"index":"not_analyzed"
},
"dish_name_analyzed": {
"type":"string",
"index":"analyzed",
"analyzer":"ik_max_word"
},
"dish_id": {
"type":"string",
"index":"not_analyzed"
},
"poi_id": {
"type":"long",
"index":"not_analyzed"
},
"poi_name": {
"type":"string",
"index":"not_analyzed"
}
}
}
}
1. 单条插入数据
1、实时导入(带ID)
curl -XPUT 'host:port/table_name/type/id?pretty' -d ' { "field": "value" }'
type:类型,可理解为二级表名。
id:该条数据的唯一id
2、实时导入(不带ID)
curl -XPOST 'host:port/table_name/type?pretty' -d ' { "field": "value" }'
不带 id 时使用 “POST”,系统将自动随机生成一个唯一id
导入时如果想更新已存在的
curl -XPOST 'host:port/table_name/type?pretty' -d '
{ “doc”:{ "field": "value" } "detect_noop":false//无视是否修改,强制合并到现有的文档 }'
批量导入:
导入文件需要如下格式:
{ "index": { "_index": "table_name", "_type": "type_name", "_id": "id1" }}
{ "field1": "value1", "field2": "value2", ……}
{ "index": { "_index": "table_name", "_type": "type_name", "_id": "id2" }}
{ "field1": "value1", "field2": "value2", ……}
_id 是可以指定为具体的field的比如下面这样(要保证值是唯一的)
{ "index": { "_index": "table_name", "_type": "type_name", "_id": "value1" }}
{ "field1": "value1", "field2": "value2", ……}
然后用以下接口导入数据:
curl -XPOST 'host:port/_bulk' --data-binary @import.json
import.json 为导入文件名。
也可在此链接中指定表名或type名,这样就不用在导入文件中指定:
curl -XPOST 'host:port/table_name/type_name/_bulk' --data-binary @import.json
***注意每个 import.json 文件不能过大,最好在10M左右,大文件可以分割为小文件并行导入***
2. 创建自定义分词和同义词配置
1:在elasticsearch-x.x.x/config目录下新建同义词文件synonym.txt。
其中,synonym.txt 编码格式为’utf-8’,内容建议为空。
2:创建索引
host:port/dishtag
{
"settings": {
"number_of_shards":3,
"number_of_replicas":1,
"index": {
"analysis": {
"analyzer": {
"by_smart": {
"type":"custom",
"tokenizer":"ik_smart",
"filter": [
"by_sfr"
]
},
"by_max_word": {
"type":"custom",
"tokenizer":"ik_max_word",
"filter": [
"by_sfr"
]
}
},
"filter": {
"by_sfr": {
"type":"synonym",
"synonyms_path":"analysis/synonym.txt"
}
}
}
}
},
"mappings": {
"dishtag": {
"_all": {
"enabled":false
},
"properties": {
"id": {
"type":"string",
"index":"not_analyzed"
},
"tag_id": {
"type":"string",
"index":"not_analyzed"
},
"tag_name": {
"type":"string",
"index":"not_analyzed",
"search_analyzer":"by_max_word",
"analyzer":"by_smart"
}
}
}
}
}
3:添加同义词
4:测试分词后同义词是否生效
host:port/dstag/_analyze?analyzer=by_max_word&pretty&text=鱼
这种结果就是同义词配置已经生效
小结
- 同义词字典或是IK用户自定义词典更新,必须每次重启elasticsearch才有效。
- 同义词词对是必须能被完成切分的词语。