Spring Data Elasticsearch 5.x新客户端下如何用NativeQuery实现聚合查询

Spring Data Elasticsearch 5.x 移除了旧的NativeSearchQueryBuilder和NativeSearchQuery，改用NativeQuery作为原生查询的载体，以下是从4.x版本代码迁移到5.x的实现：

迁移后的核心代码 #

protected <T> SearchHits<T> searchHitsForAggregation(QueryBuilder queryBuilder,
        AbstractAggregationBuilder<?> aggregationBuilder, Pageable pageable) {

    NativeQuery.NativeQueryBuilder queryBuilderNative = NativeQuery.builder()
            .withSearchType(SearchType.DEFAULT)
            .withQuery(queryBuilder)
            .withAggregation(aggregationBuilder); // 单个聚合用withAggregation，多个聚合用withAggregations传入集合

    if (pageable == null) {
        queryBuilderNative.withMaxResults(0); // 无需返回文档时设置结果数为0，避免无效数据返回
    } else {
        queryBuilderNative.withPageable(pageable);
    }

    NativeQuery nativeQuery = queryBuilderNative.build();
    return operations.search(nativeQuery, indexClass(), indexCoordinates());
}

关键变化说明 #

• 移除NativeSearchQueryBuilder和NativeSearchQuery，替换为NativeQuery.builder()构建原生查询
• 聚合添加：单个聚合调用withAggregation()方法，多个聚合可传入集合调用withAggregations(Collection<? extends AggregationBuilder>)
• 适配新客户端：5.x默认适配ElasticsearchClient（替代原RestHighLevelClient），升级依赖版本后无需额外修改客户端配置
• 聚合结果解析逻辑不变：仍通过SearchHits.getAggregations()获取聚合结果，原有解析代码可直接复用

内容的提问来源于stack exchange，提问作者user352290