elasticsearch - 使用涉及空间的正则表达式进行搜索

Question

I want to perform searching using regular expression involving whitespace in elasticsearch. I have already set my field to not_analyzed. And it's mapping is just like

"type1": {
   "properties": {
      "field1": {
         "type": "string",
         "index": "not_analyzed",
         "store": true
      }
   }
}

And I input two value for test,

"field1":"XXX YYY ZZZ"
"field1":"XXX ZZZ YYY"

And i do some case using regex query /XXX YYY/
(I want to use this query to find record1 but not record2)

{
    "query": {
        "query_string": {
           "query": "/XXX YYY/"
        }
    }
}

But it return 0 results.

However if I search without using regex (without the forward slash '/'), both record1 and record2 are returned.

Is that in elasticsearch, i cannot search using regex query involving space?

Answer 1

What you need is a ''term'' query that doesn't tokenise the search query by breaking it down into smaller parts. More about the term query here: https://www.elastic.co/guide/en/elasticsearch/reference/2.0/query-dsl-term-query.html

There's a special breed of term queries that allows you to use regexes called regexp queries. That should match any whitespaces as well: https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-regexp-query.html

Answer 2

You can keep using your query string , but your regexp is just missing a tiny part, ie the .* at the end. If you run that you'll get the single result you expect.

{
    "query": {
        "query_string": {
           "query": "/XXX YYY.*/"
        }
    }
}

Answer 3

You can use regexp queries to achieve this. Mind you, the query performance may be slow. The below query will search for all documents in which the value of field1 contains "XXX YYY".

POST <index_name>/type1/_search
{
   "query": {
      "regexp": {
         "field1": ".*XXX YYY.*"
      }
   }
}