Elasticsearch - How to search for multiple words in one string - elasticsearch

Question

I'm having issues getting the elasticsearch results i need.
My mappings look like this:
"mappings": {
"product": {
"_meta": {
"model": "App\\Entity\\Product"
},
"dynamic_date_formats": [],
"properties": {
"articleNameSearch": {
"type": "text",
"analyzer": "my_analyzer"
},
"articleNumberSearch": {
"type": "text",
"fielddata": true
},
"brand": {
"type": "nested",
"properties": {
"name": {
"type": "text"
}
}
}
}
}
},
My settings:
"settings": {
"index": {
"number_of_shards": "5",
"provided_name": "my_index",
"creation_date": "1572252785482",
"analysis": {
"filter": {
"standard": {
"type": "standard"
}
},
"analyzer": {
"my_analyzer": {
"filter": [
"standard"
],
"type": "custom",
"tokenizer": "lowercase"
}
}
},
"number_of_replicas": "1",
"uuid": "bwmc7NZ9RXqB1lpQ3e8HTQ",
"version": {
"created": "5060399"
}
}
}
The data inside:
"hits": [
{
"_index": "my_index",
"_type": "product",
"_id": "14",
"_score": 1.0,
"_source": {
"articleNumberSearch": "5003xx843",
"articleNameSearch": "this is a test string",
"brand": {
"name": "Brand name"
}
}
},
Currently the PHP code for the query looks like this (this does not return correct records):
$searchQuery = new BoolQuery();
$formattedQuery = "*" . str_replace(['.', '|'], '', trim(mb_strtolower($query))) . "*";
/**
* Test NGRAM analyzer
*/
$matchQuery = new Query\MultiMatch();
$matchQuery->setFields([
'articleNumberSearch',
'articleNameSearch',
]);
$matchQuery->setQuery($formattedQuery);
$searchQuery->addMust($matchQuery);
/**
* Nested query
*/
$nestedQuery = new Nested();
$nestedQuery->setPath('brand');
$nestedQuery->setQuery(
new Match('brand.name', 'Brand name')
);
$searchQuery->addMust($nestedQuery);
I'm creating and auto-complete search field, where you can search articleNumberSearch and articleNameSearch while brand name is always a fixed value.
I want to be able to search for example:
500 will find this hit, because 500 is in the articleNumberSearch.
But also be able to search:
this is string
Couple questions:
Which query do i need to use?
Am i using the right analyzer?
Is my anaylizer correctly configured?

Answer 1

You should create an ngram type tokenizer.
The ngram tokenizer first breaks text down into words whenever it encounters one of a list of specified characters.
Something like that:
"analysis": {
"analyzer": {
"autocomplete": {
"filter": [
"lowercase"
],
"type": "custom",
"tokenizer": "my_tokenizer"
}
},
"tokenizer": {
"my_tokenizer": {
"token_chars": [
"letter",
"digit",
"symbol",
"punctuation"
],
"min_gram": "1",
"type": "ngram",
"max_gram": "2"
}
}
}
NGram Tokenizer