Spring Data Elasticsearch 4.4.2聚合查询升级问题求助

针对升级后遇到的两处问题，直接给出适配方案和修改说明：

1. 替换addAggregation为withAggregations #

Spring Data Elasticsearch 4.4.x中，NativeSearchQueryBuilder的addAggregation方法被移除，改用withAggregations方法，该方法支持传入单个或多个聚合构建器，单个聚合场景下直接传入目标构建器即可。

2. 聚合结果的获取方式调整 #

新版本中SearchHits.getAggregations()返回AggregationsContainer类型，不再支持直接通过名称获取聚合对象，需要通过指定聚合类型的get方法来精准获取对应聚合结果。

适配后的完整代码 #

SearchSourceBuilder builder = this.getQuery(filter, false);
String aggregationName = "group_by_author_id";

TermsAggregationBuilder aggregationBuilders =
    AggregationBuilders.terms(aggregationName).field("authors");

// 改动1：用withAggregations替换原addAggregation方法
var query =
    new NativeSearchQueryBuilder()
        .withQuery(builder.query())
        .withAggregations(aggregationBuilders)
        .build();

var result = elasticsearchOperations.search(query, EsBook.class, ALIAS_COORDS);

if (!result.hasAggregations()) {
  throw new IllegalStateException("No aggregations found after query with aggregations!");
}

// 改动2：通过指定类型的get方法获取Terms聚合
Terms groupById = result.getAggregations().get(Terms.class, aggregationName);
var buckets = groupById.getBuckets();

Map<Long, Integer> booksCount = new HashMap<>();
buckets.forEach(
    bucket ->
        booksCount.put(
            bucket.getKeyAsNumber().longValue(), Math.toIntExact(bucket.getDocCount())));

return booksCount;

补充：动态遍历聚合的处理方式 #

如果需要兼容多种聚合类型或动态匹配聚合，也可以通过遍历所有聚合来筛选目标对象：

for (Aggregation aggregation : result.getAggregations().aggregations()) {
    if (aggregationName.equals(aggregation.getName()) && aggregation instanceof Terms) {
        Terms groupById = (Terms) aggregation;
        // 后续桶数据处理逻辑
        break;
    }
}

内容的提问来源于stack exchange，提问作者Sandeep Nair