使用 Whitespace Tokeniser 的弹性搜索空白分析器和客户分析器

Question

I have a custom analyser 'default' in elastic search indexing template with tokeniser as "whitespace". Can I use elastic search in-build Whitespace analyser for this purpose as my default analyser with Whitespace tokeniser and in-build Whitespace analyser will do the same task. In general, which is better to use? Will there be any performance impact?

"analysis": {
   "analyzer": {
        "default": {
          "tokenizer": "whitespace"
        }
    }
}

Answer 1

Both the whitespace tokenizer and whitespace analyzer are built-in in elasticsearch

GET /_analyze
{
  "analyzer" : "whitespace",
  "text" : "multi grain bread"
}

Following tokens are generated

{
  "tokens": [
    {
      "token": "multi",
      "start_offset": 0,
      "end_offset": 6,
      "type": "word",
      "position": 0
    },
    {
      "token": "grain",
      "start_offset": 7,
      "end_offset": 12,
      "type": "word",
      "position": 1
    },
    {
      "token": "bread",
      "start_offset": 13,
      "end_offset": 18,
      "type": "word",
      "position": 2
    }
  ]
}

You can use any of them if you only want to break the text when any whitespace comes. However, when you need to modify the whitespace , you should use the whitespace tokenizer and the filters you want to add. This is because you can not modify the whitespace analyzer.