Unable to search a query with symbols in elasticsearch

Question

I have been trying to match a query using the elasticsearch python client but I am unable to match it even after using escape characters and setting up some custom analyzers and mapping them. I want to search using & and its not giving any response.

from elasticsearch import Elasticsearch

es = Elasticsearch([{'host': 'localhost', 'port': 9200}])


doc1 = {
    'name': 'numb',
    'band': 'linkin_park',
    'year': '2006'
}

doc2 = {
    'name': 'Powerless &',
    'band': 'linkin_park',
    'year': '2006'
}
doc3 = {
    'name': 'Crawling !',
    'band': 'linkin_park',
    'year': '2006'
    }

doc =[doc1, doc2, doc3]
'''
create_index = {
    "settings": {
        "analysis": {
            "analyzer": {
                "my_analyzer": {
                    "type": "custom",
                    "filter": [
                        "lowercase"
                    ],
                    "tokenizer": "whitespace"
                }
            }
        }
    }
}

es.indices.create(index="idx_temp", body=create_index)
'''
for i in range(3):
    es.index(index="idx_temp", doc_type='_doc', id=i, body=doc[i])


my_mapping = {
  "properties": {
      "name": {
          "type": "text",
          "fields": {
              "keyword": {
                  "type": "keyword",
                  'ignore_above': 256
              }
          },
          "analyzer": "my_analyzer"
          "search_analyzer": "my_analyzer"
      },
      "band": {
          "type": "text",
          "fields": {
              "keyword": {
                  "type": "keyword",
                  "ignore_above": 256
              }
          },
          "analyzer": "my_analyzer"
          "search_analyzer": "my_analyzer"
      },
      "year": {
          "type": "text",
          "fields": {
              "keyword": {
                  "type": "keyword",
                  "ignore_above": 256
              }
          },
          "analyzer": "my_analyzer"
          "search_analyzer": "my_analyzer"
      }
  }
}

es.indices.put_mapping(index='idx_temp', body=my_mapping, doc_type='_doc', include_type_name=True)

res = es.search(index='idx_temp', body={
    "query": {
        "match": {
            "name": {
                "query": "powerless &",
                "fuzziness": 3

            }
        }
    }
})

for hit in res['hits']['hits']:
    print(hit['_source'])

The expected output was 'name': 'Poweeerless &', but i got 0 hits and no value returned.

Answer 1

So I have fixed the problem by adding another field

 "search_quote_analyzer": "my_analyzer"

to the settings field after

"analyzer": "my_analyzer"
"search_analyzer": "my_analyzer"

And then I'm getting my output by searching with & in the query as

'name': 'Poweeerless &'

Answer 2

I just tried it using your index settings, mapping, and query and was able to get the results. Below are 2 different things which I did.

1. Escape the special char &, when I was trying to index the doc using ES REST API directly, using below the body in postman:

{ "content": "Powerless \&" }

Then ES gave me the Unrecognized character escape '&' exception and even Postman, popular REST client was also giving me warning about not a proper string.

Then I changed above payload to below and was able to index the doc:

{
    "content": "Powerless \\&" :-> Notice I added a another `\` to escape the `&`
}

2. I changed the query to use the same field, which was having the value &, in your case it is name field, not the content field., As match query is analyzed and uses the same analyzer which is used for indexing time. And was able to get the result.

PS: I also verified your analyzer using _analyze api and it's generating the below tokens for text Powerless \\&

{
    "tokens": [
        {
            "token": "powerless",
            "start_offset": 0,
            "end_offset": 9,
            "type": "word",
            "position": 0
        },
        {
            "token": "\\&",
            "start_offset": 10,
            "end_offset": 12,
            "type": "word",
            "position": 1
        }
    ]
}