MySQL-如何自动执行从最近日期到前一天减去指标的视图查询&；最新数据的日期戳

我有一个每日API调用，它为我提供一个唯一ID的生命周期数据，并在每天午夜为数据批打上日期戳。这些数据被附加到MySQL中的数据库中，但我需要以某种方式对其进行转换，这样我就可以获得每个度量的唯一id的每日数据

MySQL中的我的表：

所需输出：

CREATE VIEW `facebook_insights` AS  
SELECT  
  t1.id  
, t1.timestamp  
, t1.message  
, t1.posted  
, t1.permalink_url  
, t1.caption  
, t1.link  
, t1.type  
, t1.post_impressions - t2.post_impressions as Impressions  
, t1.post_impressions_organic - t2.post_impressions_organic as Post_Impressions_Organic  
, t1.post_impressions_paid - t2.post_impressions_paid as Post_Impressions_Paid  
, t1.post_engaged_users - t2.post_engaged_users as Post_Engaged_Users  
, t1.post_consumptions - t2.post_consumptions as Post_Consumptions  
, t1.post_negative_feedback - t2.post_negative_feedback as 
Post_Negative_Feedback  
, t1.post_negative_feedback_unique - t2.Post_Negative_Feedback_Unique as 
Post_Negative_Feedback_Unique  
, t1.post_impressions_fan - t2.post_impressions_fan as Post_Impressions_Fan  
, t1.post_impressions_fan_paid - t2.post_impressions_fan_paid as 
Post_Impressions_Fan_Paid  
, t1.post_engaged_fan - t2.Post_Engaged_Fan as Post_Engaged_Fan  
, t1.post_video_complete_views_organic - 
t2.post_video_complete_views_organic as Post_Video_Complete_Views_Organic  
, t1.post_video_complete_views_paid - t2.post_video_complete_views_paid as 
Post_Video_Complete_Views_Paid  
, t1.post_video_views_10s - t2.post_video_views_10s as Post_Video_Views_10s  
, t1.post_video_views_10s_unique - t2.post_video_views_10s_unique as 
Post_Video_Views_10s_Unique  
, t1.post_video_views_organic - t2.post_video_views_organic as 
Post_Video_Views_Organic  
, t1.post_video_views_paid - t2.post_video_views_paid as 
Post_Video_Views_Paid  
, t1.post_video_views_clicked_to_play - t2.post_video_views_clicked_to_play 
as Post_Video_Views_Clicked_to_Play  

FROM  
unpaid_media.facebook_insight t1  
JOIN unpaid_media.facebook_insight t2  
ON t1.id = t2.id   
and t1.timestamp  = t2.timestamp + INTERVAL 1 DAY  

对于id 3，我将前一天（2017年7月3日）和最近日期（2017年7月4日）的差异作为其各自的指标（印象、点击次数），并将该行项目的时间戳为2017年7月4日。我需要按id处理这种类型的转换，以获得多个id和指标。谢谢大家!

根据Jacob的反馈更新我的查询：

CREATE VIEW `facebook_insights` AS  
SELECT  
  t1.id  
, t1.timestamp  
, t1.message  
, t1.posted  
, t1.permalink_url  
, t1.caption  
, t1.link  
, t1.type  
, t1.post_impressions - t2.post_impressions as Impressions  
, t1.post_impressions_organic - t2.post_impressions_organic as Post_Impressions_Organic  
, t1.post_impressions_paid - t2.post_impressions_paid as Post_Impressions_Paid  
, t1.post_engaged_users - t2.post_engaged_users as Post_Engaged_Users  
, t1.post_consumptions - t2.post_consumptions as Post_Consumptions  
, t1.post_negative_feedback - t2.post_negative_feedback as 
Post_Negative_Feedback  
, t1.post_negative_feedback_unique - t2.Post_Negative_Feedback_Unique as 
Post_Negative_Feedback_Unique  
, t1.post_impressions_fan - t2.post_impressions_fan as Post_Impressions_Fan  
, t1.post_impressions_fan_paid - t2.post_impressions_fan_paid as 
Post_Impressions_Fan_Paid  
, t1.post_engaged_fan - t2.Post_Engaged_Fan as Post_Engaged_Fan  
, t1.post_video_complete_views_organic - 
t2.post_video_complete_views_organic as Post_Video_Complete_Views_Organic  
, t1.post_video_complete_views_paid - t2.post_video_complete_views_paid as 
Post_Video_Complete_Views_Paid  
, t1.post_video_views_10s - t2.post_video_views_10s as Post_Video_Views_10s  
, t1.post_video_views_10s_unique - t2.post_video_views_10s_unique as 
Post_Video_Views_10s_Unique  
, t1.post_video_views_organic - t2.post_video_views_organic as 
Post_Video_Views_Organic  
, t1.post_video_views_paid - t2.post_video_views_paid as 
Post_Video_Views_Paid  
, t1.post_video_views_clicked_to_play - t2.post_video_views_clicked_to_play 
as Post_Video_Views_Clicked_to_Play  

FROM  
unpaid_media.facebook_insight t1  
JOIN unpaid_media.facebook_insight t2  
ON t1.id = t2.id   
and t1.timestamp  = t2.timestamp + INTERVAL 1 DAY  

---

这里的关键是将表重新连接到自身。在下面的查询中，我将重新加入您的表，但将其设置为使时间戳成为连接上的限定符。这将加入前一天的记录，并允许您简单地减去它们

选择
t1.id
，t1.时间戳
，t1.印象-t2.印象作为印象
，t1.点击-t2.点击作为点击
从…起
表t1
联接表t2
在t1.id=t2.id上
t1.timestamp=t2.timestamp+间隔1天

看起来分组函数很有用，因为它看起来像是要按日期和ID聚合印象和点击。此外，您没有提到此数据需要任何联接（尚未），因此这是假设它位于单个表中。T-SQL中的以下内容：

USE DatabaseName
GO

SELECT T.ID, 
       (CAST(T.TimeStamp) as Date) as Date, 
       SUM(T.Impressions) as TotalImpressionsPerDay, 
       SUM(T.Clicks) as TotalClicksPerDay
FROM TableName as T
GROUP BY ID, (CAST(T.TimeStamp) as Date)
ORDER BY ID ASC

---

那看起来不太好difficult@Strawberry：这并不是很难，但不如简单地使用解析函数方便。。。很多人都很难将自己的脑袋绕在一个self-joinThank@Jacobm001上，但现在我遇到了这个错误：非法混合排序规则（拉丁语-瑞典语-ci隐式）和（utf8_general_ci可强制用于操作“=”我从中调用的表的排序规则设置为latin1-默认值collation@UlisesSotomayor：这应该可以解决stackoverflow.com/a/1008336/1142368 OP有一个按日计算的总数。他想用它来生成显示每日值的输出。@Jacobm001我明白了，区别是每天增长。奇怪的是，API会以这种方式吐出生命周期数据，而不是附加新数据。