弹性搜索：简单查询未返回预期结果

Question

I'm using Elasticsearch 2.2.0 but this also happens in 1.x versions.

One of the text fields contains the word google.com. When I try to search google, ElasticSearch doesn't return anything. But if I search google.com it returns the document that contains it.

My query is simple, something like this:

query: {
  filtered: {
    query: {
      simple_query_string: {
        query: "google"
      }
    }
  }
}

What should I do to make Elastic search return the document when i search for google?

Answer 1

I do this kind of query with "wildcard" query

https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-wildcard-query.html

{
  "query": {
    "bool": {
      "must": {
        "wildcard": {
          "fieldname": "google*" // replace the fieldname her
        }
      }
    }
  }
}

Answer 2

I suggest reading the documentation about the _all field, because the simple_query_string uses this field when none are specified, which is your case.

_all is indexed by default with the standard analyzer, which means that google.com will not be split at . and will be indexed as google.com . Which means that searching for google.com will look for this term in your index whereas in your text you have google .

Searching in Elasticsearch is not only about the query, but also about the way you index the data . It may not be as simple as SQL query with LIKE but at the same time, ES gives you way more power at insert time than SQL.

Depending on what you want from your simple_query you have multiple options to choose from:

• change the analyzer of _all field:

{
  "settings": {
    "analysis": {
      "analyzer": {
        "letter": {
          "type": "custom",
          "tokenizer": "letter",
          "filter": [
            "lowercase"
          ]
        }
      }
    }
  }, 
  "mappings": {
    "test": {
      "_all": {
        "analyzer": "letter"
      }

and the query:

  "query": {
    "simple_query_string": {
      "query": "google"
    }
  }

• change the analyzer of your field and **use that field in query_string_query :

{
  "settings": {
    "analysis": {
      "analyzer": {
        "letter": {
          "type": "custom",
          "tokenizer": "letter",
          "filter": [
            "lowercase"
          ]
        }
      }
    }
  }, 
  "mappings": {
    "test": {
      "properties": {
        "text": {
          "type": "string",
          "analyzer": "letter"
        }
      }

And the query:

  "query": {
    "simple_query_string": {
      "query": "google",
      "fields": ["text"]
    }
  }

• define your own custom _all field with your own letter analyzer and use that in simple_query_string