SQL 在 MySQL 中使用 JOIN 选择常见数据和剩余数据

Question

我有三个表：wi_district、wi_group 和 wi_training。我需要根据地区计算小组和培训。为此，我使用了以下 SQL；

• SQL1

SELECT wi_district.dst_name, COUNT(grp_id) AS group_count, MAX(grp_created_date) as grp_created_date
来自 wi_group
INNER JOIN wi_district ON wi_district.dst_id=wi_group.grp_dst_id AND wi_group.grp_deleted=0 AND wi_group.grp_type IN (3)
按 wi_district.dst_name 分组

每个区的查询计数组。同样，

• SQL2

SELECT wi_district.dst_name, COUNT(trn_id) AS training_count, MAX(trn_created_date) as trn_created_date
来自 wi_training
INNER JOIN wi_district ON wi_district.dst_id=wi_training.dst_id AND wi_training.trn_deleted=0 AND wi_training.trn_beneficiary_type IN (-1,2,8,9,10)
按 wi_district.dst_name 分组

查询计算每个地区的培训。现在我需要将所有从 SQL1 和 SQL2 得到的结果结合起来，得到结果形式为

dst_name ||组数 || grp_created_date ||培训计数 || trn_created_date

问题是每当我使用 SQL1 LEFT JOIN SQL2 时，它都会显示与 SQL1 相关的结果，而 SQL2 的结果无法获得，反之亦然。请帮我解决 MySQL 中的这个问题

Answer 1

我认为你可以加入过滤表，然后按区名分组。像这样：

SELECT dist.dst_name AS dst_name,
COUNT(grp.grp_id) AS group_count, MAX(grp.grp_created_date) AS grp_created_date,
COUNT(trn.trn_id) AS training_count, MAX(trn.trn_created_date) AS trn_created_date 
FROM wi_district AS dist
LEFT JOIN (
  SELECT dst_id, trn_id, trn_created_date
  FROM wi_training
  WHERE trn_deleted=0
    AND trn_beneficiary_type IN (-1,2,8,9,10)
) AS trn ON trn.dst_id=dist.dst_id
LEFT JOIN (
  SELECT grp_dst_id, grp_id, grp_created_date
  FROM wi_group
  WHERE grp_deleted=0
    AND grp_type IN (3)
) AS grp ON grp.grp_dst_id = dist.dst_id
GROUP BY dist.dst_name

Answer 2

为什么要使用像 dst_name 这样的名称？最好把全名写出来，几个月后你也会明白它的意思。

无论如何，这个查询应该可以解决问题

select d.dst_name as district
,      count(distinct g.grp_id) as group_count
,      max(grp_created_date) as group_created_date 
,      count(distinct trn_id) as training_count
,      max(trn_created_date) as trn_created_date 
from   wi_district d
left join wi_group g on  d.dst_id = g.grp_dst_id
                     and g.grp_deleted = 0 
                     and g.grp_type in (3)
left join wi_training t on  d.dst_id = t.dst_id
                        and t.trn_deleted = 0
                        and t.trn_beneficiary_type IN (-1,2,8,9,10)
group by d.dst_name

Answer 3

您需要从分组表中驱动它。像这样。

SELECT  D.dst_name,  COUNT(grp_id) AS group_count, MAX(grp_created_date) as grp_created_date, COUNT(trn_id) AS training_count, MAX(trn_created_date) as trn_created_date 
FROM    wi_district D
LEFT JOIN wi_group G ON D.dst_id = G.grp_dst_id AND G.grp_deleted=0 AND G.grp_type IN (3)
LEFT JOIN wi_training T ON  D.dst_id = T.dst_id AND T.trn_deleted=0 AND T.trn_beneficiary_type IN (-1,2,8,9,10)
GROUP BY wi_district.dst_name

如果您只想要任一表上存在的行，请添加一个子句：

WHERE NOT G.grp_dst_id IS NULL OR NOT D.dst_id IS NULL