Elasticsearch Java API查询无法返回期望文档结果的解决求助

背景信息 #

使用Elasticsearch时，编写了以下DSL查询：

{
  "query": {
    "bool": {
      "must": [
        {"match": {
          "companyId": 35
        }},
        {
          "multi_match": {
            "query": "movie",
            "fields": ["name", "description"]
          }
        }
      ]
    }
  }
}

基于该DSL构建的Java QueryBuilder代码：

SearchRequest searchRequest = new SearchRequest("faqs");//index name
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
                
QueryBuilder mulMatch = QueryBuilders.multiMatchQuery(name,"name", "description").type("phrase").boost(5);
QueryBuilder cId = QueryBuilders.matchQuery("companyId", companyId).boost(5);
QueryBuilder must = QueryBuilders.boolQuery().must(cId).must(mulMatch).boost(5);
        
searchSourceBuilder.query(must);
searchRequest.source(searchSourceBuilder);
System.out.println("SS Query : " + searchSourceBuilder);

return searchSourceBuilder;

生成的实际查询语句：

{"query":{"bool":{"must":[{"match":{"companyId":{"query":35,"operator":"OR","prefix_length":0,"max_expansions":50,"fuzzy_transpositions":true,"lenient":false,"zero_terms_query":"NONE","auto_generate_synonyms_phrase_query":true,"boost":5.0}}},{"multi_match":{"query":"movie","fields":["description^1.0","name^1.0"],"type":"phrase","operator":"OR","slop":0,"prefix_length":0,"max_expansions":50,"zero_terms_query":"NONE","auto_generate_synonyms_phrase_query":true,"fuzzy_transpositions":true,"boost":5.0}}],"adjust_pure_negative":true,"boost":5.0}}}

期望返回的文档格式：

{
  "Id": 1,
  "name": "Controle",
  "description": "Welcome movie",
  "visibility": "2",
  "isarticleEnabled": false,
  "companyId": 35,
  "categoryId": 0        
}

但实际API返回结果为：

{
  "suggestOnly": false,
  "fragment": false
}

尝试使用SearchResponse res = client.search(searchRequest, RequestOptions.DEFAULT);，但在ElasticsearchRestTemplate和RestHighLevelClient中均无法正常获取期望结果。

解决方案 #

1. 修正结果解析逻辑 #

当前返回的{"suggestOnly": false, "fragment": false}属于高亮或建议组件的返回片段，并非完整的搜索文档。需要检查代码中是否错误提取了高亮片段，而非从搜索命中结果中获取原始文档。

2. RestHighLevelClient正确解析示例 #

如果使用RestHighLevelClient，需从SearchResponse中提取命中的文档数据：

SearchResponse response = client.search(searchRequest, RequestOptions.DEFAULT);
SearchHit[] hits = response.getHits().getHits();
List<Map<String, Object>> resultDocs = new ArrayList<>();

for (SearchHit hit : hits) {
    // 获取文档原始数据
    Map<String, Object> source = hit.getSourceAsMap();
    resultDocs.add(source);
    
    // 如需转换为实体类，可使用ObjectMapper
    // YourEntity entity = new ObjectMapper().readValue(hit.getSourceAsString(), YourEntity.class);
}

3. ElasticsearchRestTemplate正确使用示例 #

如果使用ElasticsearchRestTemplate，需通过SearchHits提取实体对象：

// 构建SearchQuery
NativeSearchQuery searchQuery = new NativeSearchQueryBuilder()
        .withQuery(must)
        .build();

// 执行查询并解析结果
SearchHits<YourEntity> searchHits = elasticsearchRestTemplate.search(searchQuery, YourEntity.class);
List<YourEntity> resultEntities = searchHits.stream()
        .map(SearchHit::getContent)
        .collect(Collectors.toList());

4. 验证查询本身的有效性 #

先通过Kibana或curl执行生成的DSL，确认Elasticsearch中存在匹配的文档：

curl -X GET "localhost:9200/faqs/_search" -H "Content-Type: application/json" -d '
{"query":{"bool":{"must":[{"match":{"companyId":{"query":35,"operator":"OR","prefix_length":0,"max_expansions":50,"fuzzy_transpositions":true,"lenient":false,"zero_terms_query":"NONE","auto_generate_synonyms_phrase_query":true,"boost":5.0}}},{"multi_match":{"query":"movie","fields":["description^1.0","name^1.0"],"type":"phrase","operator":"OR","slop":0,"prefix_length":0,"max_expansions":50,"zero_terms_query":"NONE","auto_generate_synonyms_phrase_query":true,"fuzzy_transpositions":true,"boost":5.0}}],"adjust_pure_negative":true,"boost":5.0}}}'

• 如果该查询能返回期望文档，说明问题出在Java代码的结果解析环节；
• 如果无结果返回，需检查：
  • multi_match使用了phrase类型，需确认name或description字段中存在完整的"movie"短语（而非分词后的单个词）；
  • companyId字段的映射类型是否为数值型（integer/long），避免类型不匹配导致查询失效。

内容的提问来源于stack exchange，提问作者Parva Gurav