ES学习摘要 - 8. Aggregations

ES支持的aggregation操作可以分为四类

Bucketing

每个bucket关联了一个key以及一个条件，所有满足条件的doc都会落入对应的bucket中

Metric

对一组doc进行指标运算

Pipeline

将其他aggregations的结果作为输入再做aggregation操作

Matrix

略（试验中）

aggregations部分的语法：

"aggs" : { "<aggregation_name>" : { "<aggregation_type>" : { <aggregation_body> } [,"meta" : { [<meta_data_body>] } ]? [,"aggregations" : { [<sub_aggregation>]+ } ]? } [,"<aggregation_name_2>" : { ... } ]* }

8.1 Metrics Aggregations

Avg 单值

POST /exams/_search?size=0 { "aggs" : { "avg_grade" : { "avg" : { "field" : "grade" } } } }

Cardinality 单值

calculates an approximate count of distinct values.

计算通过HyperLogLog++算法，是粗略值

POST /sales/_search?size=0 { "aggs" : { "type_count" : { "cardinality" : { "field" : "type" } } } }

Max/Min 单值<