本篇将开始介绍Elasticsearch Bucket聚合(桶聚合)。web
Buket Aggregations(桶聚合)不像metrics Aggregations(度量聚合)那样计算字段上的度量,而是建立文档桶,每一个文件桶有效地定义一个文档集。除了bucket自己以外,bucket聚合还计算并返回“落入”每一个bucket的文档的数量。数据库
与度量聚合相反,桶聚合能够嵌套子聚合。这些子聚合将为它们的“父”桶聚合建立的桶进行聚合。数组
ES Bucket Aggregations对标关系型数据库的(group by)。微信
首先咱们来介绍桶聚合两个经常使用参数intervals、time_zone的含义。svg
一、Intervals
定义桶的间隔,其可选值以下:函数
二、Time Zone
对于日期类型,可使用time_zone来指定时区,可选值能够是相对ISO 8601 utc的相对值,例如+01:00或-08:00,也能够是时区ID,例如America/Los_Angeles。源码分析
三、Histogram Aggregation
直方图聚合,Date Histogram Aggregation是其特例。ui
动态将文档中的值按照特定的间隔构建桶,并计算落在该桶的数量,文档中的值根据以下函数进行近似匹配:.net
bucket_key = Math.floor((value - offset) / interval) * interval + offset,
其中interval必须是正小数(包含正整数),offset为[0,interval)。rest
主要支持的参数以下:
具体JAVA的示例将在Date Histogram Aggregation中详细介绍。
四、Date Histogram Aggregation
日期字段直方图聚合。
4.1 interval 取值
4.2 示例
{ "aggs" : { "sales_over_time" : { "date_histogram" : { "field" : "date", "interval" : "month" } } } }
对应的JAVA示例以下:
/** * 日期直方图聚合 */ public static void test_Date_Histogram_Aggregation() { RestHighLevelClient client = EsClient.getClient(); try { //构建日期直方图聚合 时间间隔,示例中按月统计 DateHistogramInterval interval = new DateHistogramInterval("1M"); SearchRequest searchRequest = new SearchRequest(); searchRequest.indices("aggregations_index02"); SearchSourceBuilder sourceBuilder = new SearchSourceBuilder(); AggregationBuilder aggregationBuild = AggregationBuilders.dateHistogram("createTime_histogram") .field("createTime") .dateHistogramInterval(interval) // .format("yyyy-MM-dd") // 对key的格式化 ; sourceBuilder.aggregation(aggregationBuild); sourceBuilder.size(0); sourceBuilder.query( QueryBuilders.termQuery("sellerId", 24) ); searchRequest.source(sourceBuilder); SearchResponse result = client.search(searchRequest, RequestOptions.DEFAULT); System.out.println(result); } catch (Throwable e) { e.printStackTrace(); } finally { EsClient.close(client); } }
对应的返回值:
{ ... //省略常规响应 "aggregations":{ "date_histogram#createTime_histogram":{ "buckets":[ "key_as_string":"2015-12-01 00:00:00", "key":1448928000000, "doc_count":6 }, { "key_as_string":"2016-01-01 00:00:00", "key":1451606400000, "doc_count":4 } ] } } }
其相应的参数已在上面详述,在此不重复介绍。
4.3 Date Histogram聚合支持的经常使用参数
除Histogram Aggregation罗列的参数后,还额外支持以下参数:
"aggregations":{ "date_histogram#createTime_histogram":{ "buckets":{ "2015-12-01 00:00:00":{ "key_as_string":"2015-12-01 00:00:00", "key":1448928000000, "doc_count":6 }, "2016-01-01 00:00:00":{ "key_as_string":"2016-01-01 00:00:00", "key":1451606400000, "doc_count":4 } } } } }
五、Date Range Aggregation
日期范围聚合,每一个范围定义[from,to),from,to可支持date mesh格式。
其使用示例以下,其余与 Date Histogram相似。
/** * 日期范围聚合 */ public static void test_Date_range_Aggregation() { RestHighLevelClient client = EsClient.getClient(); try { //构建日期直方图聚合 时间间隔,示例中按月统计 SearchRequest searchRequest = new SearchRequest(); searchRequest.indices("aggregations_index02"); SearchSourceBuilder sourceBuilder = new SearchSourceBuilder(); AggregationBuilder aggregationBuild = AggregationBuilders.dateRange("createTime_date_range") .field("createTime") .format("yyyy-MM-dd") .addRange("quarter_01", "2016-01", "2016-03") .addRange("quarter_02", "2016-03", "2016-06") .addRange("quarter_03", "2016-06", "2016-09") .addRange("quarter_04", "2016-09", "2016-12") // .format("yyyy-MM-dd") // 对key的格式化 ; sourceBuilder.aggregation(aggregationBuild); sourceBuilder.size(0); sourceBuilder.query( QueryBuilders.termQuery("sellerId", 24) ); searchRequest.source(sourceBuilder); SearchResponse result = client.search(searchRequest, RequestOptions.DEFAULT); System.out.println(result); } catch (Throwable e) { e.printStackTrace(); } finally { EsClient.close(client); } }
六、Filter Aggregation
聚合中支持首先根据过滤上下文对全部文档进行刷选,而后再进行聚合计算,例如:
POST /sales/_search?size=0 { "aggs" : { "t_shirts" : { "filter" : { "term": { "type": "t-shirt" } }, "aggs" : { "avg_price" : { "avg" : { "field" : "price" } } } } } }
其对应的JAVA代码以下:
/** * 日期范围聚合 */ public static void test_filter_Aggregation() { RestHighLevelClient client = EsClient.getClient(); try { //构建日期直方图聚合 时间间隔,示例中按月统计 SearchRequest searchRequest = new SearchRequest(); searchRequest.indices("aggregations_index02"); SearchSourceBuilder sourceBuilder = new SearchSourceBuilder(); AggregationBuilder aggregationBuild = AggregationBuilders.filter("t_shirts", QueryBuilders.termQuery("status", "1")) .subAggregation(AggregationBuilders.avg("avg").field("num")) ; sourceBuilder.aggregation(aggregationBuild); sourceBuilder.size(0); sourceBuilder.query( QueryBuilders.termQuery("sellerId", 24) ); searchRequest.source(sourceBuilder); SearchResponse result = client.search(searchRequest, RequestOptions.DEFAULT); System.out.println(result); } catch (Throwable e) { e.printStackTrace(); } finally { EsClient.close(client); } }
其返回结果以下:
{ ... //省略 "aggregations":{ "filter#t_shirts":{ "doc_count":2, "avg#avg":{ "value":1 } } } }
{
… //省略
“aggregations”:{
“filter#t_shirts”:{
“doc_count”:2,
“avg#avg”:{
“value”:1
}
}
}
}
七、Filters Aggregation
定义一个多桶聚合,其中每一个桶与一个过滤器相关联。每一个bucket将收集与其关联过滤器匹配的全部文档。
public static void test_filters_aggregation() { RestHighLevelClient client = EsClient.getClient(); try { //构建日期直方图聚合 时间间隔,示例中按月统计 SearchRequest searchRequest = new SearchRequest(); searchRequest.indices("aggregations_index02"); SearchSourceBuilder sourceBuilder = new SearchSourceBuilder(); AggregationBuilder aggregationBuild = AggregationBuilders.filters("create_filters", QueryBuilders.termQuery("status", 1), QueryBuilders.termQuery("buyerId", 1)) .subAggregation(AggregationBuilders.avg("avg").field("num")) ; sourceBuilder.aggregation(aggregationBuild); sourceBuilder.size(0); sourceBuilder.query( QueryBuilders.termQuery("sellerId", 24) ); searchRequest.source(sourceBuilder); SearchResponse result = client.search(searchRequest, RequestOptions.DEFAULT); System.out.println(result); } catch (Throwable e) { e.printStackTrace(); } finally { EsClient.close(client); } }
其返回结果:
{ ... // 省略 "aggregations":{ "filters#create_filters":{ "buckets":[ { "doc_count":2, "avg#avg":{ "value":1 } }, { "doc_count":0, "avg#avg":{ "value":null } } ] } } }
舒适提示,每个filter表明一个桶(聚合)。
八、Global Aggregation
全局聚合,会忽略全部的查询条件,具体从下述例子进行说明:
POST /sales/_search?size=0 { "query" : { "match" : { "type" : "t-shirt" } }, "aggs" : { "all_products" : { "global" : {}, "aggs" : { "avg_price" : { "avg" : { "field" : "price" } } } }, "t_shirts": { "avg" : { "field" : "price" } } } }
其聚合的文档集不是匹配该查询的文档"query" : {“match” : { “type” : “t-shirt” } },而是针对全部的文档进行聚合。
对应的JAVA实例以下:
public static void test_global_aggregation() { RestHighLevelClient client = EsClient.getClient(); try { //构建日期直方图聚合 时间间隔,示例中按月统计 SearchRequest searchRequest = new SearchRequest(); searchRequest.indices("aggregations_index02"); SearchSourceBuilder sourceBuilder = new SearchSourceBuilder(); AggregationBuilder aggregationBuild = AggregationBuilders.global("all_producers") .subAggregation(AggregationBuilders .avg("num_avg_aggregation") .field("num")) ; sourceBuilder.aggregation(aggregationBuild); sourceBuilder.size(0); sourceBuilder.query( QueryBuilders.termQuery("sellerId", 24) ); searchRequest.source(sourceBuilder); SearchResponse result = client.search(searchRequest, RequestOptions.DEFAULT); System.out.println(result); } catch (Throwable e) { e.printStackTrace(); } finally { EsClient.close(client); } }
对应的返回值以下:
{ "took":151, "timed_out":false, "_shards":{ "total":5, "successful":5, "skipped":0, "failed":0 }, "hits":{ "total":39, // @1 "max_score":0, "hits":[ ] }, "aggregations":{ "global#all_producers":{ "doc_count":1286, // @2 "avg#num_avg_aggregation":{ "value":1.3157076205287714 } } } }
结果@1:表示符合查询条件的总个数。
结构@2:表示参与聚合的文档数量,等于当前库中文档总数。
九、IP Range Aggregation
ip类型特有的范围聚合,与其余聚合使用相似,就不重复介绍了。
十、Missing Aggregation
统计缺乏某个字段的文档个数。
JAVA示例以下:
AggregationBuilder aggregationBuild = AggregationBuilders.missing("missing_num_count") .field("num");
十一、Range Aggregation
基于多桶值源的聚合,容许用户定义一组范围——每一个范围表示一个桶。在聚合过程当中,将根据每一个bucket范围和相关/匹配文档的“bucket”检查从每一个文档中提取的值。注意,此聚合包含from值,并排除每一个范围的to值。
GET /_search { "aggs" : { "price_ranges" : { "range" : { "field" : "price", "ranges" : [ { "to" : 100.0 }, { "from" : 100.0, "to" : 200.0 }, { "from" : 200.0 } ] } } } }
对应的JAVA示例以下:
public static void test_range_aggregation() { RestHighLevelClient client = EsClient.getClient(); try { //构建日期直方图聚合 时间间隔,示例中按月统计 SearchRequest searchRequest = new SearchRequest(); searchRequest.indices("aggregations_index02"); SearchSourceBuilder sourceBuilder = new SearchSourceBuilder(); AggregationBuilder aggregationBuild = AggregationBuilders.range("num_range_aggregation") .field("num") .addRange(0, 5) .addRange(5,10) .addUnboundedFrom(10) ; sourceBuilder.aggregation(aggregationBuild); sourceBuilder.size(0); sourceBuilder.query( QueryBuilders.termQuery("sellerId", 24) ); searchRequest.source(sourceBuilder); SearchResponse result = client.search(searchRequest, RequestOptions.DEFAULT); System.out.println(result); } catch (Throwable e) { e.printStackTrace(); } finally { EsClient.close(client); } }
其返回结果以下:
{ // 省略 "aggregations":{ "range#num_range_aggregation":{ "buckets":[ { "key":"0.0-5.0", "from":0, "to":5, "doc_count":38 }, { "key":"5.0-10.0", "from":5, "to":10, "doc_count":0 }, { "key":"10.0-*", "from":10, "doc_count":1 } ] } } }
Range Aggregations支持嵌套聚合,使用subAggregations来支持嵌套聚合,根据官网示例以下:
GET /_search { "aggs" : { "price_ranges" : { "range" : { // @1 "field" : "price", "ranges" : [ { "to" : 100 }, { "from" : 100, "to" : 200 }, { "from" : 200 } ] }, "aggs" : { // @2 "price_stats" : { "stats" : { "field" : "price" } } } } } }
首先经过@1定义范围聚合,而后对每一个桶中 的文档再执行子聚合@2,其返回结果以下:
{ ... "aggregations": { "price_ranges": { "buckets": [ { "key": "*-100.0", "to": 100.0, "doc_count": 2, "price_stats": { "count": 2, "min": 10.0, "max": 50.0, "avg": 30.0, "sum": 60.0 } }, { "key": "100.0-200.0", "from": 100.0, "to": 200.0, "doc_count": 2, "price_stats": { "count": 2, "min": 150.0, "max": 175.0, "avg": 162.5, "sum": 325.0 } }, { "key": "200.0-*", "from": 200.0, "doc_count": 3, "price_stats": { "count": 3, "min": 200.0, "max": 200.0, "avg": 200.0, "sum": 600.0 } } ] } } }
本文详细介绍了ES 桶聚合,并给出JAVA示例,下一篇将重点关注ES桶聚合之term聚合。
欢迎加笔者微信号(dingwpmz),加群探讨,笔者优质专栏目录:
一、源码分析RocketMQ专栏(40篇+)
二、源码分析Sentinel专栏(12篇+)
三、源码分析Dubbo专栏(28篇+)
四、源码分析Mybatis专栏
五、源码分析Netty专栏(18篇+)
六、源码分析JUC专栏
七、源码分析Elasticjob专栏
八、Elasticsearch专栏(20篇+)
九、源码分析MyCat专栏