盐城网站建设镇江搜索优化技巧

1. 聚合可以实现对文档数据的统计，分析，运算，聚合常见有三类（聚合的值一定不能是text类型的）：

   桶（Bucket）聚合：用来对文档做分组。

   度量（Metric）聚合：用以计算一些值，比方说最大值，最小值，平均值等。

   管道（pipeline）聚合：其它聚合的结果为基础进行聚合。

   参与聚合的字段类型：keyword，数值，日期，布尔。

1. DSL实现Bucket聚合

   lasticsearch 的 Bucket 聚合（桶聚合）是将文档分组到 "桶" 中的强大工具，类似于 SQL 中的GROUP BY。每个桶关联一个条件，符合条件的文档会被分到对应的桶中。

   Terms聚合

   • 场景：统计博客文章中每个标签的文档数量。

   • GET /blog/_search
     {"size": 0,  // 不返回原始文档，只返回聚合结果"aggs": {"tags": {"terms": {"field": "tags.keyword",  // 使用keyword类型避免分词"size": 10,  // 返回前10个最常见的标签"order": {"_count": "desc"  // 按文档数量降序排序}}}}
     }
     结果示例
     {"aggregations": {"tags": {"doc_count_error_upper_bound": 0,"sum_other_doc_count": 15,"buckets": [{"key": "elasticsearch","doc_count": 25},{"key": "java","doc_count": 18}]}}
     }
     //大多数 Bucket 聚合默认按文档数量（_count）降序排序。
     //按文档数量排序DSL演示
     GET /products/_search
     {"size": 0,"aggs": {"by_category": {"terms": {"field": "category.keyword","order": {"_count": "desc"  // 按文档数量降序（默认）}}}}
     }
     //结果演示
     {"aggregations": {"by_category": {"buckets": [{ "key": "electronics", "doc_count": 120 },{ "key": "clothing", "doc_count": 80 },{ "key": "books", "doc_count": 50 }]}}
     }
     //场景：只对价格大于 100 的商品进行类别聚合
     //DSL示例
     GET /products/_search
     {"query": {"range": {"price": {"gt": 100}}},"size": 0,"aggs": {"by_category": {"terms": {"field": "category.keyword"}}}
     }
     //结果显示
     {"aggregations": {"by_category": {"buckets": [{"key": "electronics","doc_count": 100,"expensive_products": {"doc_count": 75,  // 价格>100的电子产品数量"count": {"value": 75}}}]}}
     }

   • aggs代表聚合，与query同级，此时query的作用是限定聚合的的文档范围

   • 聚合必须的三要素

     • 聚合名称

     • 聚合类型

     • 聚合字段

   • 聚合可配置的属性有：size：指定聚合结果数量，order指定聚合结果排序方式，field指定聚合字段。

2. DSL实现Metric聚合

   计算所有产品的平均价格

   GET /products/_search
   {"size": 0,  // 不返回原始文档"aggs": {"avg_price": {"avg": {"field": "price"}}}
   }
   //结果显示
   {"aggregations": {"avg_price": {"value": 125.5  // 平均价格}}
   }

   嵌套聚合metric聚合的组合使用

   //按类别分组，计算每个类别的平均价格、最高价格和最低价格。
   GET /products/_search
   {"size": 0,"aggs": {"by_category": {"terms": {"field": "category.keyword"},"aggs": {"avg_price": { "avg": { "field": "price" } },"max_price": { "max": { "field": "price" } },"min_price": { "min": { "field": "price" } },"price_stats": { "stats": { "field": "price" } }}}}
   }

   在java中进行聚合

   import org.elasticsearch.action.search.SearchRequest;
   import org.elasticsearch.action.search.SearchResponse;
   import org.elasticsearch.client.RequestOptions;
   import org.elasticsearch.client.RestHighLevelClient;
   import org.elasticsearch.index.query.QueryBuilders;
   import org.elasticsearch.search.aggregations.AggregationBuilders;
   import org.elasticsearch.search.aggregations.bucket.filter.Filter;
   import org.elasticsearch.search.aggregations.bucket.terms.Terms;
   import org.elasticsearch.search.builder.SearchSourceBuilder;
   ​
   import java.io.IOException;
   ​
   public class FilterAggregationExample {private final RestHighLevelClient client;public FilterAggregationExample(RestHighLevelClient client) {this.client = client;}public void filterAggregation() throws IOException {SearchRequest searchRequest = new SearchRequest("products");SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();// 构建按类别分组的聚合，在每个类别中过滤价格>100的商品searchSourceBuilder.aggregation(AggregationBuilders.terms("by_category").field("category.keyword").subAggregation(AggregationBuilders.filter("expensive_products",QueryBuilders.rangeQuery("price").gt(100)).subAggregation(AggregationBuilders.valueCount("count").field("id"))));searchRequest.source(searchSourceBuilder);SearchResponse response = client.search(searchRequest, RequestOptions.DEFAULT);// 处理聚合结果Terms byCategory = response.getAggregations().get("by_category");for (Terms.Bucket bucket : byCategory.getBuckets()) {String category = bucket.getKeyAsString();long totalCount = bucket.getDocCount();Filter expensiveProducts = bucket.getAggregations().get("expensive_products");long expensiveCount = expensiveProducts.getDocCount();System.out.println("Category: " + category + ", Total: " + totalCount + ", Expensive: " + expensiveCount);}}
   }