Fuzzy Matching Email or Telephone Using Elasticsearch

Elasticsearch offers built-in capabilities for fuzzy matching of email addresses and telephone numbers.

Email Matching

To match email addresses ending with a specific domain (e.g., @gmail.com):

{
    "query": {
        "term": {
            "email": ".*@gmail.com"
        }
    }
}

Or, to match emails containing a specific string:

{
    "query": {
        "match": {
            "email": {
                "query": "sales@*",
                "operator": "and"
            }
        }
    }
}

Telephone Matching

For fuzzy matching of telephone numbers, you can use the following pattern:

{
    "query": {
        "prefix": {
            "tel": "136*"
        }
    }
}

This will match all phone numbers starting with "136".

Performance Optimization

To improve performance for fuzzy matching, consider using custom analyzers that leverage n-gram or edge n-gram token filters. These filters break down the text into smaller tokens, making it easier for Elasticsearch to perform fuzzy matching.

Email Analyzer Configuration:

{
  "settings": {
    "analysis": {
      "analyzer": {
        "email_analyzer": {
          "type": "custom",
          "tokenizer": "standard",
          "filter": [
            "lowercase",
            "name_ngram_filter",
            "trim"
          ]
        }
      },
      "filter": {
        "name_ngram_filter": {
          "type": "ngram",
          "min_gram": "3",
          "max_gram": "20"
        }
      }
    }
  }
}

Telephone Analyzer Configuration:

{
  "settings": {
    "analysis": {
      "analyzer": {
        "phone_analyzer": {
          "type": "custom",
          "char_filter": [
            "digit_only"
          ],
          "tokenizer": "digit_edge_ngram_tokenizer",
          "filter": [
            "trim"
          ]
        }
      },
      "char_filter": {
        "digit_only": {
          "type": "pattern_replace",
          "pattern": "\\D ",
          "replacement": ""
        }
      },
      "tokenizer": {
        "digit_edge_ngram_tokenizer": {
          "type": "edgeNGram",
          "min_gram": "3",
          "max_gram": "15",
          "token_chars": [
            "digit"
          ]
        }
      }
    }
  }
}