Phrase_prefix search issue with single character

[dipanshu_sharma]

I have a data like 'Dipanshu S', 'Mohit Sharma' and 'Mohit Shakya' in my database and when i do a ES search query like :-

"multi_match": {
"query": "Mohit S",
"type": "phrase_prefix",
"fields": [ "name", "nickname" ]
}

It gives me correct result 'Mohit Sharma' and 'Mohit Shakya' as it returns results with exact matching of first term 'Mohit' and returning all words starting with 'S' but when i do same search with query

"multi_match": {
"query": "Dipanshu S",
"type": "phrase_prefix",
"fields": [ "name", "nickname" ]
}

:"Dipanshu S" then it doesn't return any results because the word 'S' is not a word starting with character 'S' but its a complete word in itself in the data 'Dipanshu S'.

So how to manipulate query to get results with single character words?

[spinscale (Alexander Reelsen)]

can you please share a fully reproducible example including index creation, document indexing and mapping...

This example works for me, so something in your setup seems to be different

DELETE test

PUT test/_doc/1?refresh
{
  "name" : "Dipanshu S"
}

GET test/_search
{
  "query": {
    "multi_match": {
      "query": "Dipanshu S",
      "type": "phrase_prefix",
      "fields": [
        "name"
      ]
    }
  }
}

Thanks!

--Alex

[dipanshu_sharma]

I'm sharing my mappings and settings(analyzers) configs.
Mapping:

 "name": {
                        "type": "keyword",
                        "fields": {
                            "caseInsensitive": {
                                "type": "text",
                                "analyzer": "lowercase_keyword_analyzer"
                            },
                            "fulltext": {
                                "type": "text",
                                "analyzer": "english_standard_edgengram_analyzer",
                                "search_analyzer": "english_standard_analyzer"
                            }
                        },
                        "normalizer": "lowercase_normalizer"
                    }

Analyzer:

 "analyzer": {
                        "numeric_edgengram_analyzer": {
                            "filter": [
                                "edgengram_filter"
                            ],
                            "char_filter": [
                                "numeric_filter"
                            ],
                            "type": "custom",
                            "tokenizer": "keyword"
                        },
                        "lowercase_keyword_analyzer": {
                            "filter": [
                                "lowercase"
                            ],
                            "type": "custom",
                            "tokenizer": "keyword"
                        },
                        "numeric_keyword_analyzer": {
                            "type": "custom",
                            "char_filter": [
                                "numeric_filter"
                            ],
                            "tokenizer": "keyword"
                        },
                        "english_standard_analyzer": {
                            "filter": [
                                "lowercase"
                            ],
                            "type": "keyword"
                        },
                        "english_standard_edgengram_analyzer": {
                            "filter": [
                                "lowercase",
                                "edgengram_filter"
                            ],
                            "type": "custom",
                            "tokenizer": "english_standard_tokenizer"
                        }
                    }

And you can test with some different data like "Dipanshu S Gupta", "Dipanshu Sharma".

[spinscale (Alexander Reelsen)]

can you please share the full commands of index creation + mapping in the console tools or in a curl compatible fashion? The above snippets means, everyone reading your ask for support needs to convert this in working snippets, which again is quite tedious.

Thank you!

[dipanshu_sharma]

hey, it's little bit difficult for me to share that code but just wants to know that does the query "Dipanshu S" gives the result "Dipanshu S Sharma" ? I mean does the prefix "S" includes the result "Dipanshu S Sharma" as "S" is not a word starting from "S" but it's a complete word in itself. And did you got my question?