Elasticsearch multi_match query not searching in beginning of string-CodePudding

I am matching the phone numbers and ssn that start with '40'. For ssn I am getting the correct matching count. For the phone number, I am not getting the correct matching count, as the phone number contains hyphens '-' in between the phone number. Example: '403-517-2323'.

When I search for a phone number that starts with '40' it includes the phone number that has '40' in between after the hyphen. Example: '222-401-8120' and '823-093-4012'.

How can I exclude matching in between and match only at the start of the phone number?

Below is the query I am trying>

GET emp_details_1_1/_msearch 
{"index": "emp_details_1_1"}
{"_source":[],"size":0,"min_score":1,"query":{"multi_match":{"query":"40","fields":["ssn"],"type":"phrase_prefix"}}}
{"index": "emp_details_1_1"}
{"_source":[],"size":0,"min_score":1,"query":{"multi_match":{"query":"21","fields":["phone"],"type":"phrase_prefix"}}}

CodePudding user response：

As I don't have the exact index mapping and settings, I am guessing you are using the analyzer that is breaking your phone number on _, if its default analyzer(standard) then it does the same thing.

POST _analyze
{
    "text" : "403-517-2323",
    "analyzer": "standard"
}

Tokens generated

{
    "tokens": [
        {
            "token": "403",
            "start_offset": 0,
            "end_offset": 3,
            "type": "<NUM>",
            "position": 0
        },
        {
            "token": "517",
            "start_offset": 4,
            "end_offset": 7,
            "type": "<NUM>",
            "position": 1
        },
        {
            "token": "2323",
            "start_offset": 8,
            "end_offset": 12,
            "type": "<NUM>",
            "position": 2
        }
    ]
}

In order to fix that, you should use the prefix query on .keyword subfield field(if its generated in your mapping) or create a field of type keyword that stores phone number in your index.