Spring Boot中用RestHighLevelClient按条件查询Elasticsearch索引数据的实现

前置检查 #

• 确认索引es_52_companydetails_index中companyaddress字段映射为nested类型，否则嵌套数组查询无法精准过滤单个子元素，会默认返回数组全部内容，无法满足第二个查询场景的要求

场景1：按company_name查询，返回该公司所有company_name/address3组合 #

@Autowired
private RestHighLevelClient restHighLevelClient;

public List<String> queryByCompanyName(String companyName) throws IOException {
    SearchRequest searchRequest = new SearchRequest("es_52_companydetails_index");
    SearchSourceBuilder sourceBuilder = new SearchSourceBuilder();
    // 精准匹配company_name，字段为text类型需加.keyword后缀，keyword类型可直接写字段名
    sourceBuilder.query(QueryBuilders.termQuery("company_name.keyword", companyName));
    // 仅返回业务需要的字段，减少IO消耗
    sourceBuilder.fetchSource(new String[]{"company_name", "companyaddress.address3"}, null);
    searchRequest.source(sourceBuilder);

    SearchResponse response = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT);
    List<String> result = new ArrayList<>();
    for (SearchHit hit : response.getHits().getHits()) {
        Map<String, Object> sourceMap = hit.getSourceAsMap();
        String name = (String) sourceMap.get("company_name");
        List<Map<String, Object>> addressList = (List<Map<String, Object>>) sourceMap.get("companyaddress");
        for (Map<String, Object> address : addressList) {
            result.add(name + "/" + address.get("address3"));
        }
    }
    return result;
}

场景2：按address3查询，仅返回匹配的company_name/address3组合 #

public List<String> queryByAddress3(String address3) throws IOException {
    SearchRequest searchRequest = new SearchRequest("es_52_companydetails_index");
    SearchSourceBuilder sourceBuilder = new SearchSourceBuilder();
    // 构建nested嵌套查询
    NestedQueryBuilder nestedQuery = QueryBuilders.nestedQuery("companyaddress",
            QueryBuilders.termQuery("companyaddress.address3.keyword", address3),
            ScoreMode.None);
    sourceBuilder.query(nestedQuery);
    // 主查询仅返回公司名称，匹配的地址从inner_hits中获取
    sourceBuilder.fetchSource(new String[]{"company_name"}, null);
    // 开启inner_hits只返回命中的嵌套子对象，而非整个数组
    nestedQuery.innerHit(new InnerHitBuilder("matched_address"));
    searchRequest.source(sourceBuilder);

    SearchResponse response = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT);
    List<String> result = new ArrayList<>();
    for (SearchHit hit : response.getHits().getHits()) {
        String name = (String) hit.getSourceAsMap().get("company_name");
        // 从inner_hits中提取匹配的address3
        SearchHits innerHits = hit.getInnerHits().get("matched_address");
        for (SearchHit innerHit : innerHits.getHits()) {
            String matchedAddr = (String) innerHit.getSourceAsMap().get("address3");
            result.add(name + "/" + matchedAddr);
        }
    }
    return result;
}

补充说明 #

• 上述代码默认字段为text+keyword双类型，若你的索引中字段本身就是keyword类型，去掉查询条件中的.keyword后缀即可
• IOException的捕获和处理逻辑可根据项目自身的异常规范调整

内容的提问来源于stack exchange，提问作者Dorayya