应该在 bigquery 中提取不匹配的记录

Question

所需的结果集：

在下面的查询中，我正在获取按旧表和源表分组的记录，如果值匹配，则比较结果，然后我的标志列将填充匹配 但我希望输出中也有不匹配的记录（附件中的 HTC 产品不存在于源表中，因此该记录未填充）我的结果中也需要 HTC - 请帮助

查询框：

with legacy as (select product,sum(premium) as premium_l,reportdate
from Tab_legacy
group by product,reportdate),

source as (select product,sum(premium) as premium_s,reportdate
from Tab_src
group by product,reportdate)

select legacy.product,legacy.premium_l,source.premium_s,
case when legacy.premium_l = source.premium_s then 'match' else 'unmatch' end as match_flg,
round((legacy.premium_l - source.premium_s )/source.premium_s,2) as Diff,
source.reportdate
from legacy join source
on legacy.reportdate = source.reportdate
and legacy.product = source.product

Answer 1

你可以使用full join:

with legacy as (
      select product, sum(premium) as premium_l, reportdate
      from Tab_legacy
      group by product, reportdate
     ),
     source as (
      select product, sum(premium) as premium_s, reportdate
      from Tab_src
      group by product, reportdate
     )
select product, reportdate, legacy.premium_l, source.premium_s,
       (case when legacy.premium_l = source.premium_s
             then 'match' else 'unmatch'
        end) as match_flg,
       round((legacy.premium_l - source.premium_s )/source.premium_s,2) as Diff,
source.reportdate
from legacy full join
     source
     using (reportdate, product);

注意变化：

• full join 而不是 join。
• using 而不是 on。这使得在select 中包含键值变得更容易（否则您需要coalesce() 逻辑）。
• reportdate 在select 中。它是join 的一部分，因此您应该包含它。