问题背景

_elasticsearch 的 keyword/raw 是对字段值的逐字存储的，就是没有做分词，所以对英文字母大小写敏感。当我们使用wildcard 做模糊查询时就会遇到查询结果集不全或者查询不到结果集_

解决方案

_我们需要在索引文档的时候对raw/keyword做下特殊里，下面是具体实现步骤_

• 在setting中设置分析器

PUT /test_raw

{

"settings": {

"analysis": {

"filter": {

"my_synonym": {

"type": "synonym",

"synonyms_path": "synonym.dic"

}

},

"analyzer": {

"pinyin_analyzer": {

"tokenizer": "my_pinyin"

},

"first_letter_pinyin_analyzer": {

"tokenizer": "my_first_letter_pinyin"

},

"my_ik_max_word": {

"tokenizer": "ik_max_word",

"filter": [

"my_synonym"

]

},

"my_ik_smart": {

"tokenizer": "ik_smart",

"filter": [

"my_synonym"

]

}

},

"normalizer": {

"lowercase": {

"type": "custom",

"filter": [

"lowercase"

]

}

},

"tokenizer": {

"my_pinyin": {

"type": "pinyin",

"keep_first_letter": true,

"keep_separate_first_letter": false,

"keep_full_pinyin": false,

"keep_original": false,

"lowercase": true

},

"my_first_letter_pinyin": {

"type": "pinyin",

"keep_first_letter": false,

"keep_separate_first_letter": false,

"keep_joined_full_pinyin": true,

"keep_none_chinese": true,

"keep_none_chinese_in_joined_full_pinyin": true,

"none_chinese_pinyin_tokenize": false,

"keep_full_pinyin": false,

"keep_original": false,

"limit_first_letter_length": 50,

"trim_whitespace": false,

"lowercase": true

}

}

},

"number_of_shards": 1,

"number_of_replicas": 0,

"max_result_window": "1000000"

},

"mappings": {

"dynamic_templates": [

{

"strings": {

"match_mapping_type": "string",

"mapping": {

"type": "text",

"store": true,

"fields": {

"spy": {

"type": "text",

"store": false,

"fielddata": "true",

"analyzer": "pinyin_analyzer"

},

"ik": {

"type": "text",

"store": false,

"analyzer": "ik_max_word",

"search_analyzer": "ik_smart"

},

"syno_ik": {

"type": "text",

"store": false,

"analyzer": "my_ik_max_word",

"search_analyzer": "ik_smart"

},

"raw": {

"type": "keyword",

"normalizer": "lowercase"

},

"py": {

"type": "text",

"store": false,

"fielddata": "true",

"analyzer": "first_letter_pinyin_analyzer"

}

}

}

}

}

]

}

}

上面给raw设置了 "normalizer": "lowercase" 

• 添加数据

POST test_raw/_doc/1

{

"name":"Low case Hello",

"title":"我爱北京天安门ttTT Hello",

"content":"世界那么大nibu Hello",

"completion":123

}

POST test_raw/_doc/2

{

"name":"Low case hello",

"title":"我爱北京天安门ttTT hello",

"content":"世界那么大nibu hello",

"completion":123

}

上面添加了两个文档，分别有Hello 和 hello

• ik分词查询效果

GET test_raw/_search

{

"query": {

"multi_match": {

"query": "Hello",

"fields": ["title.ik","name.ik"]

}

}

}

结果集：

{

"took" : 1,

"timed_out" : false,

"_shards" : {

"total" : 1,

"successful" : 1,

"skipped" : 0,

"failed" : 0

},

"hits" : {

"total" : {

"value" : 2,

"relation" : "eq"

},

"max_score" : 0.18232156,

"hits" : [

{

"_index" : "test_raw",

"_type" : "_doc",

"_id" : "1",

"_score" : 0.18232156,

"_source" : {

"name" : "Low case Hello",

"title" : "我爱北京天安门ttTT Hello",

"content" : "世界那么大nibu Hello",

"completion" : 123

}

},

{

"_index" : "test_raw",

"_type" : "_doc",

"_id" : "2",

"_score" : 0.18232156,

"_source" : {

"name" : "Low case hello",

"title" : "我爱北京天安门ttTT hello",

"content" : "世界那么大nibu hello",

"completion" : 123

}

}

]

}

}

• wildcard 效果

GET test_raw/_search

{

"query": {

"wildcard": {

"name.raw": {

"value": "*Hello*"

}

}

}

}

结果集：

{

"took" : 0,

"timed_out" : false,

"_shards" : {

"total" : 1,

"successful" : 1,

"skipped" : 0,

"failed" : 0

},

"hits" : {

"total" : {

"value" : 2,

"relation" : "eq"

},

"max_score" : 0.18232156,

"hits" : [

{

"_index" : "test_raw",

"_type" : "_doc",

"_id" : "1",

"_score" : 0.18232156,

"_source" : {

"name" : "Low case Hello",

"title" : "我爱北京天安门ttTT Hello",

"content" : "世界那么大nibu Hello",

"completion" : 123

}

},

{

"_index" : "test_raw",

"_type" : "_doc",

"_id" : "2",

"_score" : 0.18232156,

"_source" : {

"name" : "Low case hello",

"title" : "我爱北京天安门ttTT hello",

"content" : "世界那么大nibu hello",

"completion" : 123

}

}

]

}

}

可以看到满足了我们预期；