如何在索引时使用弹性搜索验证数据

Question

我需要防止某些具有“null”（null 作为字符串）和""（空字符串）等值的字段在 Elasticsearch 中被编入索引，即我应该能够获取不包含这些值的文档_source 字段中的字段。

索引时映射中是否需要任何设置，例如在字段上使用自定义分析器？

P.S:- 我使用的是 elasticsearch 7.6.1

我尝试了以下答案，但这就是它不起作用的原因-

{  "settings": {
"number_of_shards": "5",
"analysis": {
  "normalizer": {
    "my_normalizer": {
      "char_filter": [
        {
          "type": "mapping",
          "mappings": [
            "null =>",
            "\"\"\" =>"
          ]
        }
      ],
      "filter": [
        "uppercase"
      ],
      "type": "custom"
    }
  }
},
"number_of_replicas": "1"}}

响应错误 - 序列化设置中只允许使用值列表

即使我尝试了以下设置，我也没有得到预期的结果：

{  "settings": {
"number_of_shards": "5",
"analysis": {
  "char_filter": {
    "my_filter": {
      "type": "mapping",
      "mappings": [
        "null =>",
        "\"\"\" =>"
      ]
    }
  },
  "normalizer": {
    "my_normalizer": {
      "char_filter": [
        "my_filter"
      ],
      "filter": [
        "uppercase"
      ],
      "type": "custom"
    }
  }
},
"number_of_replicas": "1"}}

请求 - GET indexname/_analyze

{"normalizer":"my_normalizer","text":"null"}

响应 -

{
"tokens": [
    {
        "token": "",
        "start_offset": 4,
        "end_offset": 4,
        "type": "word",
        "position": 0
    }
]

}

预期响应 -

{
"tokens": []
}

Answer 1

在您的分析器定义中使用mapping char filter，您可以实现它，下面是工作示例。

分析 API

{
  "tokenizer": "standard",
  "char_filter": [
    {
      "type": "mapping",
      "mappings": [
        "null =>",
        "\"\"\" =>"
      ]
    }
  ],
  "text": "null" or "" --> note this
}

并返回令牌

{
    "tokens": []
}