Char_filter and lowercase not working for search_analyzer

Elastic Stack Elasticsearch
1

I've been experimenting with many different iterations of the below but haven't been able to match all the cases I'd like to.

My objective is to match the following cases:

  • "7-eleven" or "7eleven" to the text "7-Eleven"
  • "At&t" or "ATT" to "AT&T"

Mapping:

{
  "settings": {
    "analysis": {
      "char_filter" : {
        "remove_special_chars" : {
          "type" : "mapping",
          "mappings" : [
            "-=>",
            "&=>"
          ]
        }
      },
      "tokenizer": {
        "autocomplete": {
          "type": "edge_ngram",
          "min_gram": 1,
          "max_gram": 15,
          "token_chars": [
            "letter",
            "digit"
          ]
        }
      },
      "analyzer": {
        "autocomplete": {
          "tokenizer": "autocomplete",
          "filter": [
            "lowercase"
          ],
          "char_filter" : ["remove_special_chars"]
        },
        "autocomplete_search": {
          "tokenizer": "standard",
          "filter": [
            "lowercase"
          ],
          "char_filter" : ["remove_special_chars"]
        }
      }
    }
  },
  "mappings": {
    "vendors": {
      "_all": {},
      "properties": {
        "searchable": {
          "type": "text",
          "analyzer": "autocomplete",
          "search_analyzer": "autocomplete_search"
        }
      }
    }
  }
}

Add Document:

put my_test_index/vendors/1
{
    "searchable": "7-Eleven"
}

Querying the documents

{
  "sort" : [
    "_score"
  ],
  "query": {
    "bool": {
      "should": [
          { "match" :
            { "searchable" :
              {
                "query" : "7-eleven",
                "operator" : "and"
              }
            }
          }
        ]
    }
  }
}

The above works for a query of 7eleven but not 7-eleven. Is something wrong with my syntax?Why is the query not going through the search analyzer?

2

OMG, there was some code deep within our project which was escaping the hyphen character.
The code above works as expected.

3

glad you found it, for future reference you might want to take a look at the explain API as well as the explain parameter in the search bodry.

1 Like
4

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.