elasticsearch时间柱状图聚合java实现

需求：统计每天有多少条数据,以及某字段的平均值。

1. 使用DateHistogramAggregationBuilder按天统计，再嵌入aggs取平均值。

如果要对结果进行排序，在 AggregationBuilders 加上 .order(Histogram.Order.COUNT_DESC) 

1
2
3
4
5
6
7
8
9
10
11
AvgAggregationBuilder avgAggregationBuilder = AggregationBuilders
        .avg("avg_aggsName")
        .field("fieldName");
        
DateHistogramAggregationBuilder dateHistogramAggregationBuilder = AggregationBuilders
        .dateHistogram("aggsName")
        .field("fieldName") //可以是time
        .dateHistogramInterval(DateHistogramInterval.DAY)
        .format("yyyy-MM-dd")
        .minDocCount(0L)
        .subAggregation(avgAggregationBuilder);
2. 如果再新增一条需求：只统计近一个月的数据。

那么就加一条filter，过滤后再聚合。

1
2
3
4
5
6
7
8
9
10
11
12
//日期限制
		QueryBuilder rangeBuilder = QueryBuilders
            .rangeQuery(aggsName)
            .format("yyyy-MM-dd'T'HH:mm:ss.SSS'Z'")
            .gte(timeRange.get("startTime").toString())
            .lte(timeRange.get("endTime").toString());
            
//        过滤后聚合 .missing(0)
        QueryBuilder queryBuilder = QueryBuilders
                .boolQuery()
                .filter(rangeBuilder);
3. 获取当天/一周内／一个月内起始时间
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
private static Map<String, String> getTime(String period) {
       Map<String, String> timeRange = new HashMap<>();
       SimpleDateFormat simpleDateFormat = new SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ss.SSS'Z'");
       Calendar calendar = Calendar.getInstance();
       timeRange.put("endTime", simpleDateFormat.format(calendar.getTime()));
       switch (period) {
           case "week": {
               calendar.add(Calendar.DATE, -7);
           }
           break;
           case "month": {
               calendar.add(Calendar.MONTH, -1);
           }
           break;
           default: break;
       }
       calendar.set(Calendar.HOUR_OF_DAY, 0);
       calendar.set(Calendar.MINUTE, 0);
       calendar.set(Calendar.SECOND, 0);
       calendar.set(Calendar.MILLISECOND, 0);
       timeRange.put("startTime", simpleDateFormat.format(calendar.getTime()));
       return timeRange;
   }
4. 整合查询语句
1
2
3
4
5
searchSourceBuilder
              .query(queryBuilder)
              .aggregation(dateHistogramAggregationBuilder);
      String query = searchSourceBuilder.toString();

需求：获取type/index下最新一条记录

首先用 matchAll 匹配所有数据，再根据时间排序。最终获取 getFirstHit，得到最新记录。（字段中最好包含time字段）

1
2
3
4
5
6
QueryBuilder queryBuilder = QueryBuilders
              .matchAllQuery();
      
      searchSourceBuilder
              .query(queryBuilder)
              .sort("time", SortOrder.DESC);

输出结果

1. 获取 hits 数据：
1
2
3
4
5
List<SearchResult.Hit<TESTCLASS, Void>> hits = result.getHits(TESTCLASS.class);
      List<TESTCLASS> userList = new ArrayList<>();
      for (SearchResult.Hit<TESTCLASS, Void> hit : hits) {
          userList.add(hit.source);
      }
2. 获取 aggs 里的 buckets 数据：
1
MetricAggregation jsonAggs = searchResult.getAggregations();

日期柱状图

1
2
3
4
5
Map<String,Object> map = new HashMap<>();
      DateHistogramAggregation histogram = jsonAggs.getDateHistogramAggregation("aggsName");
      for (DateHistogramAggregation.DateHistogram entry : histogram.getBuckets()) {
          map.put(entry.getTimeAsString(), entry.getCount());
      }

求平均值

1
2
AvgAggregation avg = jsonAggs.getAvgAggregation("avg_aggsName");
result.put("avg_aggsName", Math.ceil(avg.getAvg()));