我有 EmailEvent 表。这是存储打开和点击的 STI 表(通过使用 type 属性)。我现在有这段代码，它提供有关有多少电子邮件使用收件人地址(email_address 字段)的信息:
EmailEvent.where("type = 'OpenEmailEvent' OR type = 'ClickEmailEvent'")
.where(email_id: email_ids).happened_between_dates(start_date, end_date)
.select("email_address, count(distinct email_id) as emails_count")
.group(:email_address).reorder(nil)
它生成这样的代码:
SELECT email_address, count(distinct email_id) as emails_count
FROM "email_events"
WHERE "email_events"."deleted_at" IS NULL
AND (type = 'OpenEmailEvent' OR type = 'ClickEmailEvent')
AND "email_events"."email_id" IN (85487, 75554, 85489, 77184, 78562, 75587, 82879, 85535, 85534)
AND (created_at between '2017-02-28 22:52:01.000000' AND '2017-03-29 05:59:59.000000')
GROUP BY "email_events"."email_address"
我需要提供有关每个 email_address 有多少打开和点击的信息。我不能使用 count(id) 因为它会给我打开和点击，所以我需要以某种方式使用 type 属性但我无法弄清楚如何它应该工作。有人可以给我建议我应该在这里做什么吗？
最佳答案
替换这个
count(distinct email_id) as emails_count
在 SQLServer 中使用 case 或 IIF 之类的东西
sum(case when type = 'OpenEmailEvent' then 1 else 0 end) [Opens],
sum(case when type = 'ClickEmailEvent' then 1 else 0 end) [Click],
关于sql - 如何根据另一个字段计算字段值？，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/43081670/