Elasticsearch 7.9.1中ngram tokenizer的max_ngram未生效求助

问题原因 #

你的索引配置中，ngram_tokenizer的min_ngram和max_ngram参数使用了字符串类型（带引号的"2"和"40"），但Elasticsearch要求这两个参数必须是整数类型。这导致配置未被正确解析，Elasticsearch使用了Ngram Tokenizer的默认值：min_ngram=1、max_ngram=2，所以只会生成1-2字符的token。

解决步骤 #

由于Elasticsearch创建索引后无法修改分析器配置，需要删除现有索引并重新创建：

1. 删除旧索引

curl -X DELETE 'http://elasticsearch_test:9200/test'

2. 创建新索引（使用正确的整数参数）

{
  "settings": {
    "index": {
      "max_ngram_diff": 40,
      "number_of_shards": 1,
      "number_of_replicas": 1,
      "analysis": {
        "analyzer": {
          "ngram_analyzer": {
            "filter": [
              "lowercase",
              "trim"
            ],
            "type": "custom",
            "tokenizer": "ngram_tokenizer"
          },
          "search_analyzer": {
            "filter": [
              "lowercase",
              "trim"
            ],
            "type": "custom",
            "tokenizer": "keyword"
          }
        },
        "tokenizer": {
          "ngram_tokenizer": {
            "type": "ngram",
            "max_ngram": 40,
            "min_ngram": 2
          }
        }
      }
    }
  }
}

执行创建命令：

curl -X PUT 'http://elasticsearch_test:9200/test' \
--header 'Content-Type: application/json' \
--data '@上述配置文件路径'

3. 重新测试分析器
   使用你原来的测试命令：

curl --location --request GET 'http://elasticsearch_test:9200/test/_analyze?pretty=true' \
--header 'Content-Type: application/json' \
--data '{
  "analyzer": "ngram_analyzer",
  "text": "cod_1"
}'

预期结果 #

此时会生成所有长度从2到5（输入文本总长度为5）的ngram，示例返回如下：

{
  "tokens" : [
    {
      "token" : "co",
      "start_offset" : 0,
      "end_offset" : 2,
      "type" : "word",
      "position" : 0
    },
    {
      "token" : "cod",
      "start_offset" : 0,
      "end_offset" : 3,
      "type" : "word",
      "position" : 1
    },
    {
      "token" : "cod_",
      "start_offset" : 0,
      "end_offset" : 4,
      "type" : "word",
      "position" : 2
    },
    {
      "token" : "cod_1",
      "start_offset" : 0,
      "end_offset" : 5,
      "type" : "word",
      "position" : 3
    },
    {
      "token" : "od",
      "start_offset" : 1,
      "end_offset" : 3,
      "type" : "word",
      "position" : 4
    },
    {
      "token" : "od_",
      "start_offset" : 1,
      "end_offset" : 4,
      "type" : "word",
      "position" : 5
    },
    {
      "token" : "od_1",
      "start_offset" : 1,
      "end_offset" : 5,
      "type" : "word",
      "position" : 6
    },
    {
      "token" : "d_",
      "start_offset" : 2,
      "end_offset" : 4,
      "type" : "word",
      "position" : 7
    },
    {
      "token" : "d_1",
      "start_offset" : 2,
      "end_offset" : 5,
      "type" : "word",
      "position" : 8
    },
    {
      "token" : "_1",
      "start_offset" : 3,
      "end_offset" : 5,
      "type" : "word",
      "position" : 9
    }
  ]
}

内容的提问来源于stack exchange，提问作者Olego