Elasticsearch嵌套查询must和mustNot_elasticsearch must

场景：在bug关联固件的时候将bug的数据放到固件的数据下，可以根据固件数据下是否包含bug数据查询出已关联和未关联的数据。

ES文档结构

目录

1.must查询此bug关联的固件

java代码

2.mustNot查询此bug未关联的固件

java代码

3.劫后余生       

4.闲来无事must_not跟must下的多条件关系不一样

4.1 es查询

4.2 java代码

---

---

1.must查询此bug关联的固件

{
    "query":{
        "nested":{
            "path":"s25_kernel_check",
            "query":{
                "bool":{
                    "must":[
                        {
                            "match":{
                                "s25_kernel_check.cve_id":"value"
                            }
                        },
                        {
                            "match":{
                                "s25_kernel_check.exploit-db":"value"
                            }
                        }
                    ]
                }
            }
        }
    }
}

同时满足must下的条件才可以被返回

java代码

BoolQueryBuilder must = QueryBuilders.boolQuery();
if (cvesById.getCveId()!=null){
	must.must(QueryBuilders.matchQuery("s25_kernel_check.cve_id", cvesById.getCveId()));
}
if (cvesById.getExploitDb()!=null){
	must.must(QueryBuilders.matchQuery("s25_kernel_check.exploit-db", cvesById.getExploitDb()));
}
 
HashMap<String, Object> map2 = new HashMap<>();
NestedQueryBuilder s25_kernel_check = QueryBuilders.nestedQuery("s25_kernel_check", must,ScoreMode.Max);
SearchQuery queryBuilder = new NativeSearchQueryBuilder()
		.withQuery(s25_kernel_check)
		.withFields("firmware_id")
		.build();
AggregatedPage<Cvesdd> page = template.queryForPage(queryBuilder, Cvesdd.class, new SearchResultMapper() {
	@Override
	public <T> AggregatedPage<T> mapResults(SearchResponse searchResponse, Class<T> aClass, Pageable pageable) {
		map2.put("total", searchResponse.getHits().totalHits);
		List<T> list = new ArrayList<>();
		return new AggregatedPageImpl<T>(list);
	}
});
Integer total= new Integer(String.valueOf((Long) map2.get("total")));
if (total == 0){
	total = 10;
}
NestedQueryBuilder new_s25_kernel_check = QueryBuilders.nestedQuery("s25_kernel_check", must,ScoreMode.Max);
SearchQuery new_queryBuilder = new NativeSearchQueryBuilder()
		.withQuery(new_s25_kernel_check)
		.withFields("firmware_id")
		.withPageable(PageRequest.of(0,total))
		.build();
List<Cvesdd> list = template.queryForList(new_queryBuilder, Cvesdd.class);

2.mustNot查询此bug未关联的固件

{
  "size": 200, 
  "query": {
    "bool": {
      "must_not": [
        {
          "nested": {
            "path": "s25_kernel_check",
            "query": {
              "bool": {
                "must": [
                  {
                    "match": {
                      "s25_kernel_check.cve_id": "value"
                    }
                  },
                  {
                    "match": {
                      "s25_kernel_check.exploit-db": "value"
                    }
                  }
                ]
              }
            }
          }
        }
      ]
    }
  }
}

排除同时满足must下条件的数据并返回

java代码

HashMap<String, Object> totalMap = new HashMap<>();
BoolQueryBuilder must = QueryBuilders.boolQuery();
if (cvesById.getCveId()!=null){
	must.must(QueryBuilders.matchQuery("s25_kernel_check.cve_id", cvesById.getCveId()));
}
if (cvesById.getExploitDb()!=null){
	must.must(QueryBuilders.matchQuery("s25_kernel_check.exploit-db", cvesById.getExploitDb()));
}
NestedQueryBuilder nestedQuery = QueryBuilders.nestedQuery("s25_kernel_check", must,ScoreMode.Max);
BoolQueryBuilder newBool = QueryBuilders.boolQuery();
newBool.mustNot(nestedQuery);
SearchQuery queryBuilder = new NativeSearchQueryBuilder()
		.withQuery(newBool)
		.withFields("firmware_id")
		.build();
AggregatedPage<Cvesdd> page = template.queryForPage(queryBuilder, Cvesdd.class, new SearchResultMapper() {
	@Override
	public <T> AggregatedPage<T> mapResults(SearchResponse searchResponse, Class<T> aClass, Pageable pageable) {
		totalMap.put("total", searchResponse.getHits().totalHits);
		List<T> list = new ArrayList<>();
		return new AggregatedPageImpl<T>(list);
	}
});
Integer total= new Integer(String.valueOf((Long) totalMap.get("total")));
SearchQuery new_queryBuilder = new NativeSearchQueryBuilder()
		.withQuery(newBool)
		.withFields("firmware_id")
		.withPageable(PageRequest.of(0,total))
		.build();
List<Cvesdd> list = template.queryForList(new_queryBuilder, Cvesdd.class);

查询语句跟代码的对应关系

---

3.劫后余生       

        OK，上面是满足了我想要的效果，但是发现在关联后，在es中再增加一条不同任务但是固件已经被某条bug关联过的数据，就会发现某条bug的未关联固件和已关联固件都会出现这个固件，原因是es中其实是存放的任务跟固件的关联关系，又因为业务关系，会存在多条相同固件但是不同任务的数据，关联之后创建的新任务已有固件数据，其中并未包含某条bug的数据，所以才会导致这样，那么ok，在查询未关联的逻辑之前，先查询一下已经关联的数据，不管它在哪个任务，只要所有的固件数据，有一条满足包含此条bug数据，那么我们就认为此条固件已经关联了这个bug了

4.闲来无事must_not跟must下的多条件关系不一样

        还尝试了排除嵌套查询满足或者id存在集合中的数据，gpt说这俩是and的关系，但是怎么测都是or的关系，现在es有三条id等于137的数据，其中有两条是满足嵌套查询的条件，如果是and那么应该是给我返回只是id等于137不满足嵌套查询的数据，但是并没有。。

        但是如果将must_not改成must下面的这个两个就成了and的关系，就会返回同时满足这两个条件的数据

        就很奇怪！离谱它妈给离谱开门，离谱到家了~

        那么我们来看看gpt对疑问的回答

4.1 es查询

{
  "size": 200, 
  "query": {
    "bool": {
      "must_not": [
        {
          "terms": {
            "firmware_id": [
              137
            ]
          }
        },
        {
          "nested": {
            "path": "s25_kernel_check",
            "query": {
              "bool": {
                "must": [
                  {
                    "match": {
                      "s25_kernel_check.cve_id": "CVE-2010-3848,CVE-2010-3850,CVE-2010-4073"
                    }
                  },
                  {
                    "match": {
                      "s25_kernel_check.exploit-db": "17787"
                    }
                  }
                ]
              }
            }
          }
        }
      ]
    }
  }
}

4.2 java代码

---

 最上面所提到的关联操作

//查询es中要关联的固件数据
BoolQueryBuilder boolQuery = QueryBuilders.boolQuery();
boolQuery.must(QueryBuilders.matchQuery("firmware_id",ids[i]));
SearchQuery searchQuery = new NativeSearchQueryBuilder()
		.withQuery(boolQuery)
		//指定索引
		.withIndices("information_result")
		//指定type
		.withTypes("fulldata")
		.build();
List<?> results = template.query(searchQuery, response -> {
	List<Object> list = new ArrayList<>();
	for (SearchHit hit : response.getHits()) {
		Map<String, Object> sourceAsMap = hit.getSourceAsMap();
		sourceAsMap.put("id",hit.getId());
		list.add(sourceAsMap);
	}
	return list;
});
 
//构建批量插入的集合
List<IndexQuery> queries = new ArrayList<>();
for (int i1 = 0; i1 < results.size(); i1++) {
	Map oldData = (HashMap) results.get(i1);
	ArrayList s25_kernel_check1 =new ArrayList();
	boolean key1 = oldData.containsKey("s25_kernel_check");
	if (key1){
		s25_kernel_check1 = (ArrayList) oldData.get("s25_kernel_check");
	}
	//添加的bug信息
	HashMap<String, Object> stringObjectHashMap = new HashMap<>();
	//.....添加本条新bug数据
	s25_kernel_check1.add(stringObjectHashMap);
	oldData.put("s25_kernel_check",s25_kernel_check1);
	//构建插入数据
	IndexQuery indexQuery = new IndexQuery();
	//指定id,覆盖原有数据
	indexQuery.setId(oldData.get("id").toString());
	oldData.remove("id");
	JSONObject newData=new JSONObject(oldData);
	//插入数据
	indexQuery.setSource(newData.toString());
	//索引
	indexQuery.setIndexName("information_result");
	//类型
	indexQuery.setType("fulldata");
	//3.添加到queries
	queries.add(indexQuery);
}
//4.添加数据
template.bulkIndex(queries);
queries.clear();