Creating an ElasticSearch query to search all fields and use partial matching at the same time

Question

Separately, I have partial search working and I can search all fields. But I'm unable use partial matching on a search that maps all fields.

This returns my name when I use partial matching:

GET /_search?pretty=true
{ 
    "query": {
        "match": {
            "FullName": "andon"
        }
    }
}

This search runs on all fields, but doesn't do partial matching (0 hits):

GET /_search?pretty=true
{
    "query": {
        "match": {
           "_all": "andon"
        }
    }
}

What do I need to do to combine partial matching with a search on all fields & types? I have many fields and types so I'm hoping to avoid having to specify them all manually.

If it makes any difference, I'm inserting my data from a River that connects to SQL.

EDIT: I now have the _all field in my mapping. It appears to make no difference.

PUT /contact/
{
    "settings" : {
        "index" : {
            "analysis" : {
                "analyzer" : {
                    "my_ngram_analyzer" : {
                        "tokenizer" : "my_ngram_tokenizer"
                    }
                },
                "tokenizer" : {
                    "my_ngram_tokenizer" : {
                        "type" : "nGram",
                        "min_gram" : "1",
                        "max_gram" : "10"
                    }
                }
            }
        }
    },
    "mappings": {
        "_default_": {
            "_all" : {"enabled" : true},
            "properties" : {
                "Description":{"type":"string","store":"yes", "index_analyzer": "my_ngram_analyzer"},
                "OfficePhone":{"type":"string","store":"yes"},
                "Email":{"type":"string","store":"yes", "index_analyzer": "my_ngram_analyzer"},
                "FullName":{"type":"string","store":"yes", "index_analyzer": "my_ngram_analyzer"},
                "ReportsTo":{"type":"string","store":"yes"},
                "Department":{"type":"string","store":"yes", "index_analyzer": "my_ngram_analyzer"},
                "Title":{"type":"string","store":"yes", "index_analyzer": "my_ngram_analyzer"}
            }
        }
    }
}

Answer 1

The _all-field works by taking the text for all fields, pass them through the analyzer for _all (which is standard unless overridden), then index the resulting terms.

It does not take whatever terms are made by the analyzers for the other fields and shove them all into the same field.

Thus, you'd have to override the _all-field to use your custom analyzer.

But! I don't think you'd really want to do that. You'll get a much larger index. Having _all is convenient at times, but it's usually better to control what fields are searched at search time.

So instead of doing a match on _all, you probably want to be using multi_match, specifying your fields. You can also consider copy_to if you're on 1.0.

Answer 2

I tried searching with wildcard characters(%2A) and was able to do partial searches

ex => ?q=chew%2A&size=100