如何在sql中找到两个表的值之间的差异？

Question

我有一个表格，其中包含销售和购买产品的供应商的数据。该表包含产品名称、产品价值和交易类型（“购买”或“出售”）。我想写一个查询来了解每个特定产品的损益。 表格示例：

所以我能够找到每个产品对应的值的总和。

select product_name, SUM(product_value) as sell from vendor where transaction_type="sell" group by product_name;
select product_name, SUM(product_value) as buy from vendor where transaction_type="buy" group by product_name;

现在我有了这两个子表，但是如何对这两个表进行差分操作呢？我想要一个结果表，其中包含每个产品的 product_name 和损益值（总销售成本 - 总购买成本）。

Answer 1

也许您需要一个简单的条件案例来转换您的值，然后您可以简单地减去：

select productname, 
Coalesce(Sum(case when TransactionType='sell' then productvalue end),0) -
Coalesce(Sum(case when TransactionType='buy' then productvalue end),0) ProfitLoss
from t
group by productname

Example Fiddle

Answer 2

只是猜测，但也许

with tb1 as (
  select product_name, SUM(product_value) as sell
  from vendor
  where transaction_type='sell'
  group by product_name
),
tb2 as (
  select product_name, SUM(product_value) as buy
  from vendor
  where transaction_type='buy'
  group by product_name
)
select tb1.product_name, tb1.sell, tb2.buy, (tb2.buy - tb1.sell) as profit
from tb1 left join tb2 on tb1.product_name=tb2.product_name

使用此设置：

create table vendor (product_name text, product_value int, transaction_type text)
insert into vendor values ('ABC',3196,'sell'),('XYZ',3196,'sell'),('PQR',3196,'sell'),('ABC',2365,'buy'),('ABC',2365,'buy')

查询返回

product_name  sell   buy profit
         ABC  3196  4730   1534
         PQR  3196    NA     NA
         XYZ  3196    NA     NA

---

编辑：这也适用于 MySQL，不知道为什么它不适合你。

Docker 设置：

r2@d2sb2:~$ docker pull mysql:latest
latest: Pulling from library/mysql
...
Digest: sha256:99e0989e7e3797cfbdb8d51a19d32c8d286dd8862794d01a547651a896bcf00c
Status: Downloaded newer image for mysql:latest
docker.io/library/mysql:latest

r2@d2sb2:~$ docker run --rm -p 3306:3306 --name some-mysql -e MYSQL_ROOT_PASSWORD=my-secret-pw -d mysql:latest
2e93acda94fdb386a9205ccda8df9b4f47b5f6a9bbdd159f84550418dff4a56e

然后在 SQL 中：

SHOW VARIABLES LIKE 'version';
# Variable_name  Value
#       version 8.0.26

create schema stackoverflow;
use stackoverflow;
create table vendor (product_name text, product_value int, transaction_type text)
insert into vendor values ('ABC',3196,'sell'),('XYZ',3196,'sell'),('PQR',3196,'sell'),('ABC',2365,'buy'),('ABC',2365,'buy')

with tb1 as (
  select product_name, SUM(product_value) as sell
  from vendor
  where transaction_type='sell'
  group by product_name
),
tb2 as (
  select product_name, SUM(product_value) as buy
  from vendor
  where transaction_type='buy'
  group by product_name
)
select tb1.product_name, tb1.sell, tb2.buy, (tb2.buy - tb1.sell) as profit
from tb1 left join tb2 on tb1.product_name=tb2.product_name
# product_name sell  buy profit
#          ABC 3196 4730   1534
#          XYZ 3196   NA     NA
#          PQR 3196   NA     NA

（注意：我在 R DBI 客户端中逐一运行命令，而不是在 mysql 控制台中；我怀疑除了 R 的 NA 作为空值之外它会有所不同。仅供参考。 )

由于您的实例似乎在没有HAVING 的情况下使用SUM 存在一些问题，因此该查询模式会产生相同的结果：

with tb1 as (
  select product_name, SUM(product_value) as sell
  from vendor
  group by product_name, transaction_type
  having transaction_type='sell'
),
tb2 as (
  select product_name, SUM(product_value) as buy
  from vendor
  group by product_name, transaction_type
  having transaction_type='buy'
)
select tb1.product_name, tb1.sell, tb2.buy, (tb2.buy - tb1.sell) as profit
from tb1 left join tb2 on tb1.product_name=tb2.product_name