时间戳

首页 > 解决方案 > 对数组进行分组和计数

clickhouse - 对数组进行分组和计数

问题描述

从relaxo.tracks 中选择arrayReduce('groupUniqArray', groupArray(browser))；

arrayReduce不适用于任意 lambda。有没有办法计算数组中出现元素的计数？喜欢

select groupArray(age) from customers;
:) [21, 40, 20, 20, 20, 30]
select arrayReduce('groupUniqArray', groupArray(age)) from customers;
:) [21, 40, 20, 30]
select arrayReduce('???', groupArray(age)) from customers;
:) [(21, 1), (40, 1), (20, 3), (30, 1)]

输出格式不是那么重要。我不想在这里使用 group-by/count，因为我想通过一个查询来聚合多个字段。

select 
  arrayReduce('???', groupArray(age)),
  arrayReduce('???', groupArray(job)),
  arrayReduce('???', groupArray(country))
from customers;

像这样

标签： clickhouse

解决方案

只需进行几个数组的操作：

SELECT
    groupArray(age) AS ages,
    arrayReduce('groupUniqArray', ages) AS uniqAges,
    arraySort(x -> x.1, arrayMap(x -> (x, countEqual(ages, x)), uniqAges)) AS resultAges,

    groupArray(job) AS jobs,
    arrayReduce('groupUniqArray', jobs) AS uniqJobs,
    arraySort(x -> x.1, arrayMap(x -> (x, countEqual(jobs, x)), uniqJobs)) AS resultJobs,

    groupArray(country) AS countries,
    arrayReduce('groupUniqArray', countries) AS uniqCountries,
    arraySort(x -> x.1, arrayMap(x -> (x, countEqual(countries, x)), uniqCountries)) AS resultCountries
FROM test.test4
FORMAT Vertical

推荐阅读