使用array_agg和struct后如何在列中查找值? [英] How to find a value in a column after I've used array_agg and struct?

问题描述

在我的数据库中使用以下代码后，得到以下结果:

After using the following code in my database, I got the following results:

select USERID, 
  array_agg(struct(ORDER_TIME, DELIVERY_TIME, PLATFORM) order by ORDER_TIME) STATS
from `project.dataset.table`
group by USERID 
having count(1) > 1
order by USERID

<头>

用户 ID	STATS.ORDER_TIME	STATS.DELIVERY_TIME	统计平台
011	2021-01-09 11:14:18	2021-01-09 11:44:01	手机
	2021-02-12 16:15:51	2021-02-12 17:16:51	桌面
	2021-03-30 17:23:45	2021-02-12 17:16:51	桌面
033	2021-01-01 12:30:14	2021-01-01 13:30:00	手机
	2021-04-16 23:00:45	2021-04-16 23:45:40	手机
040	2021-02-18 19:22:55	2021-02-18 20:00:05	手机
	2021-05-06 09:12:13	2021-05-06 10:00:10	桌面

但是，我只需要那些同时包含移动设备和桌面设备的寄存器.所以我需要这样的东西，结果是没有 USERID 040 的数据，因为他们只在手机上订购:

However, I only need those registers which contains both mobile AND desktop. So I need something like this, a result where there is no data for the USERID 040, because they only ordered on a mobile phone:

<头>

用户 ID	STATS.ORDER_TIME	STATS.DELIVERY_TIME	统计平台
011	2021-01-09 11:14:18	2021-01-09 11:44:01	手机
	2021-02-12 16:15:51	2021-02-12 17:16:51	桌面
	2021-03-30 17:23:45	2021-02-12 17:16:51	桌面
033	2021-01-01 12:30:14	2021-01-01 13:30:00	手机
	2021-05-06 09:12:13	2021-05-06 10:00:10	桌面

我怎么可能这样做?非常感谢！

How can I possibly do that? Thank you very much!

推荐答案

最简单"的方法是在 have 子句

The "simplest" way is to add few more conditions into having clause

select USERID, array_agg(struct(ORDER_TIME, DELIVERY_TIME, PLATFORM) order by ORDER_TIME) STATS
from `project.dataset.table`
group by USERID 
having count(1) > 1
and 'mobile' in unnest(array_agg(PLATFORM))
and 'desktop' in unnest(array_agg(PLATFORM))
order by USERID     

如果应用于您问题中的样本数据 - 输出为

if applied to sample data in your question - output is

如果您有更多此类条目要进行比较 - 您可以使用以下版本以避免重复类似的代码行

In cases when you have more such entries to compare with - you can use below version to avoid repeating similar line of code

select USERID, array_agg(struct(ORDER_TIME, DELIVERY_TIME, PLATFORM) order by ORDER_TIME) STATS
from `project.dataset.table`
group by USERID 
having count(1) > 1
and array_length(array_agg(distinct if(PLATFORM in ('mobile', 'desktop'), PLATFORM, null))) = 2
order by USERID 

这篇关于使用array_agg和struct后如何在列中查找值?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！