CODEWITHSUNDEEP
javaelasticsearch
MassimoB
MassimoB

Reputation: 77

Java Elastic Search query gives wrong results

I have a problem with elasticsearch query in Java This is my code:

BoolQueryBuilder qb = QueryBuilders.boolQuery();

for example:
if (myFilter.getTerm1() != null) {
        qb.should(QueryBuilders.termsQuery("myfield1", myFilter.getTerm1()));
    }

if (myFilter.getTerm2() != null) {
        qb.should(QueryBuilders.termsQuery("myfield2", myFilter.getTerm2()));
    }

qb.minimumShouldMatch("100%");
SearchRequest request = op.searchRequest(classTarget(MyClass.class));// this just create a new SearchRequest    
sourceBuilder.query(QueryBuilders.simpleQueryStringQuery(qb.toString()));
    sourceBuilder.from((pageNumber));
    sourceBuilder.size(pageSize);
    request.source(sourceBuilder);

The problem is that this gives me back all the result that matches with “myfield1 OR myfield2”, but I need to have all the result that matches “myfield1 AND myfield2”

For example if I have 2 documents, one with

myfield1 = “Foo” and myfield2 = “Test”

and the other with:

myfield1 = “Elastic” and myfield2 = “Search”

if I made a query with myfield1 = “Foo” and myfield2 = “Search” I expect to have no results, but I have both documents.

I also tried to use “must” instead of “should” and I tried to set the default operator to “AND” in sourceBuilder but in this case I have no results at all, also if the query should return something.

this is the print of "request"

Query request: SearchRequest{searchType=QUERY_THEN_FETCH, indices=[myIndex], indicesOptions=IndicesOptions[ignore_unavailable=false, allow_no_indices=true, expand_wildcards_open=true, expand_wildcards_closed=false, allow_aliases_to_multiple_indices=true, forbid_closed_indices=true, ignore_aliases=false], types=[data], routing='null', preference='data', requestCache=null, scroll=null, maxConcurrentShardRequests=0, batchedReduceSize=512, preFilterShardSize=128, allowPartialSearchResults=null, source={"from":0,"size":100,"query":{"simple_query_string":{"query":"{\n  \"bool\" : {\n    \"should\" : [\n      {\n        \"terms\" : {\n          \"myfield1.keyword\" : [\n            \"foo\"\n          ],\n          \"boost\" : 1.0\n        }\n      },\n      {\n        \"terms\" : {\n          \"myfield2.keyword\" : [\n            \"search\"\n          ],\n          \"boost\" : 1.0\n        }\n      }\n    ],\n    \"adjust_pure_negative\" : true,\n    \"minimum_should_match\" : \"100%\",\n    \"boost\" : 1.0\n  }\n}","flags":-1,"default_operator":"or","analyze_wildcard":false,"auto_generate_synonyms_phrase_query":true,"fuzzy_prefix_length":0,"fuzzy_max_expansions":50,"fuzzy_transpositions":true,"boost":1.0}}}}

Upvotes: 1

Views: 319

Answers (1)

Val
Val

Reputation: 217304

If you need AND logic, then you can rule out should. You need must instead and probably with a match query instead of term (as I don't think you need exact matching). Try this instead:

if (myFilter.getTerm1() != null) {
    qb.must(QueryBuilders.matchQuery("myfield1", myFilter.getTerm1()));
}

if (myFilter.getTerm2() != null) {
    qb.must(QueryBuilders.matchQuery("myfield2", myFilter.getTerm2()));
}

also remove qb.minimumShouldMatch("100%");

Also this QueryBuilders.simpleQueryStringQuery(qb.toString()) should result in an error, as you cannot serialize a bool query into a simple_query_string query, it's just not meant for that purpose.

Simply change your statement to this one:

sourceBuilder.query(qb);

If you print out your query you'll see the following, which is most probably not what you want:

{
  "from": 0,
  "size": 100,
  "query": {
    "simple_query_string": {
      "query": "{\n  \"bool\" : {\n    \"should\" : [\n      {\n        \"terms\" : {\n          \"myfield1.keyword\" : [\n            \"foo\"\n          ],\n          \"boost\" : 1.0\n        }\n      },\n      {\n        \"terms\" : {\n          \"myfield2.keyword\" : [\n            \"search\"\n          ],\n          \"boost\" : 1.0\n        }\n      }\n    ],\n    \"adjust_pure_negative\" : true,\n    \"minimum_should_match\" : \"100%\",\n    \"boost\" : 1.0\n  }\n}",
      "flags": -1,
      "default_operator": "or",
      "analyze_wildcard": false,
      "auto_generate_synonyms_phrase_query": true,
      "fuzzy_prefix_length": 0,
      "fuzzy_max_expansions": 50,
      "fuzzy_transpositions": true,
      "boost": 1
    }
  }
}

Upvotes: 1

Related Questions