Spring 带约束的引导弹性搜索查询

Question

I am pretty new to this elastic search and spring boot world, I have been trying to search for this solution from past 2-3 days, sadly I was unable to (probably because I am new to this).

I have 3 columns inside elastic search, the first being the Id, second being name (upon which I am creating autocomplete API), and the third field with set of numbers called numbers. My use case is that for a particular number in the 3rd column, I want autocomplete suggestions from 2nd column.

This I have implemented with this code and it works: repo.findByNumbersAndName(String, String)

But this does not allow me to set fuzziness, wildcardquery and other search parameter, so I tried using QueryBuilders but I am not successful in building logic (I basically tried using BooleanQueryBuilder), So if anyone can help me on this it would be really helpful!!

Spring boot version - 2.4.2

Thanks in advance.

Edit1:-

To give better understanding of my use case, let say this is what I have stored in elastic Search

Id | Name   | Numbers |
-----------------------
a1 | ashwin | 1       |
a2 | Ram    | 3       |
a3 | Kumar  | 2       |
a4 | Some   | 2       |
a5 | body   | 1       |
a6 | any    | 3       |
a7 | one    | 4       |
a8 | ashwin | 2       |

Now I should have the control to specify my query saying that for this number (let's say 1), what are the autocomplete possibilities if input for autocomplete "a". Then the program should search for possibilities that have number "1" in the numbers column only, so in this scenario there is only one output that is "ashwin".

Edit2:-

I believe my configuration and query method is same as what you have done, I will just paste what I have done. One update is that though the Number field in database contains Integers but they are stored as String datatype, would that make any difference while term Query?

This is my elastic-analyzer.json

{

  "analysis": {
    "filter": {
      "autocomplete_filter": {
        "type": "edge_ngram",
        "min_gram": 3,
        "max_gram": 30
      }
    },
    "analyzer": {
      "autocomplete_search": {
        "type": "custom",
        "tokenizer": "standard",
        "filter": [
          "lowercase",
          "stop"
          
        ]
      },
      "autocomplete_index": {
        "type": "custom",
        "tokenizer": "standard",
        "filter": [
          "lowercase",
          "autocomplete_filter",
          "stop"
        ]
      }
    }
  }
}

I am using this to generate a Query: -

BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery()
        
        .must(QueryBuilders.termQuery("Number", number))
        .must(QueryBuilders.matchQuery("Name", search_word));

The above generates the query in this format -

{
  "bool" : {
    "must" : [
      {
        "term" : {
          "Number" : {
            "value" : "1",
            "boost" : 1.0
          }
        }
      },
      {
        "match" : {
          "Name" : {
            "query" : "ash",
            "operator" : "OR",
            "prefix_length" : 0,
            "max_expansions" : 50,
            "fuzzy_transpositions" : true,
            "lenient" : false,
            "zero_terms_query" : "NONE",
            "auto_generate_synonyms_phrase_query" : true,
            "boost" : 1.0
          }
        }
      }
    ],
    "adjust_pure_negative" : true,
    "boost" : 1.0
  }

This is producing empty array, not able to understand why? The below I feel is equivalent of properties that you had mentioned in Index Mapping

@Id
@Field(type = FieldType.Text)
private String Id;

@Field(type = FieldType.Text, analyzer = "autocomplete_index", searchAnalyzer = "autocomplete_search")
private String Name;

@Field(type = FieldType.Text)
private String Number;

Thanks @ESCoder

Edit 3:-

I stumbled upon this with help from @ESCoder as well,

BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery()
        
        .must(QueryBuilders.matchQuery("Number", search_number).boost(1f).operator(Operator.AND))
        .must(QueryBuilders.matchQuery("Name", keyword).boost(0.4f));

This is the Query spring generated for ElasticSearch,

{
  "bool" : {
    "must" : [
      {
        "match" : {
          "Number" : {
            "query" : "1",
            "operator" : "AND",
            "prefix_length" : 0,
            "max_expansions" : 50,
            "fuzzy_transpositions" : true,
            "lenient" : false,
            "zero_terms_query" : "NONE",
            "auto_generate_synonyms_phrase_query" : true,
            "boost" : 1.0
          }
        }
      },
      {
        "match" : {
          "Name" : {
            "query" : "ash",
            "operator" : "OR",
            "fuzziness" : "1",
            "prefix_length" : 0,
            "max_expansions" : 50,
            "fuzzy_transpositions" : true,
            "lenient" : false,
            "zero_terms_query" : "NONE",
            "auto_generate_synonyms_phrase_query" : true,
            "boost" : 0.4
          }
        }
      }
    ],
    "adjust_pure_negative" : true,
    "boost" : 1.0
  }
}

Answer 1

Adding a working example with index data, mapping, search query and search result

Index Mapping:

{
  "settings": {
    "analysis": {
      "filter": {
        "autocomplete_filter": {
          "type": "edge_ngram",
          "min_gram": 1,
          "max_gram": 4
        }
      },
      "analyzer": {
        "autocomplete": {
          "type": "custom",
          "tokenizer": "standard",
          "filter": [
            "lowercase",
            "autocomplete_filter"
          ]
        }
      }
    }
  },
  "mappings": {
    "properties": {
      "Name": {
        "type": "text",
        "analyzer": "autocomplete",
        "search_analyzer": "standard"
      },
      "Id": {
        "type": "text"
      },
      "Numbers": {
        "type": "integer"
      }
    }
  }
}

Index Data:

{
  "Name": "ashwin",
  "Id": "a1",
  "Numbers": 1
}
{
  "Name": "ashwin",
  "Id": "a1",
  "Numbers": 2
}

Search Query:

{
  "query": {
    "bool": {
      "must": [
        {
          "match": {
            "Name": "a"
          }
        },
        {
          "term": {
            "Numbers": 1
          }
        }
      ]
    }
  }
}

Search Result:

"hits": [
      {
        "_index": "66923434",
        "_type": "_doc",
        "_id": "1",
        "_score": 1.2630212,
        "_source": {
          "Name": "ashwin",
          "Id": "a1",
          "Numbers": 1
        }
      }
    ]

Update 1:

With your current index mapping setting, the following tokens are generated for ashwin

GET /_analyze
{
  "analyzer" : "autocomplete",
  "text" : "ashwin"
}

Tokens are:

{
  "tokens": [
    {
      "token": "ash",
      "start_offset": 0,
      "end_offset": 6,
      "type": "<ALPHANUM>",
      "position": 0
    },
    {
      "token": "ashw",
      "start_offset": 0,
      "end_offset": 6,
      "type": "<ALPHANUM>",
      "position": 0
    },
    {
      "token": "ashwi",
      "start_offset": 0,
      "end_offset": 6,
      "type": "<ALPHANUM>",
      "position": 0
    },
    {
      "token": "ashwin",
      "start_offset": 0,
      "end_offset": 6,
      "type": "<ALPHANUM>",
      "position": 0
    }
  ]
}

You need to modify your index mapping, instead of "min_gram": 3 you need to change it to "min_gram": 1 .

Update 2:

You need to even change "search_analyzer" to standard .