我需要所有相关儿童记录状态已关闭的所有门票。因此,如果其中一个孩子是其他孩子,我不希望结果集中的票证。
我尝试使用antijoin patern,但我的问题是孩子们住在不同的桌子里。
请看这个http://sqlfiddle.com/#!3/febde/8 例如
t1:票 t2:relatedrecord t3:child1 t4:child1
select ticket.ticketid, rr.relatedrecordkey, rr.relatedrecordclass, c1.id, c1.status, c2.id, c2.status
from ticket
inner join relatedrecord rr on rr.recordkey = ticket.ticketid and rr.class = ticket.class
left join child1 c1 on c1.id = rr.relatedreckey and c1.class = rr.relatedrecclass
left join child2 c2 on c2.id = rr.relatedreckey and c2.class = rr.relatedrecclass
结果:
+----------+--------+-----------+-------+-----------------+---------------+--------+--------+--------+--------+
| ticketid | status | RECORDKEY | class | RELATEDRECCLASS | relatedreckey | id | status | id | status |
+----------+--------+-----------+-------+-----------------+---------------+--------+--------+--------+--------+
| 1183 | NEW | 1183 | SR | WORKORDER | 1238 | 1238 | NEW | (null) | (null) |
| 1183 | NEW | 1183 | SR | SR | 1184 | (null) | (null) | 1184 | NEW |
| 1185 | NEW | 1185 | SR | WORKORDER | 1239 | 1239 | CLOSE | (null) | (null) |
| 1185 | NEW | 1185 | SR | SR | 1186 | (null) | (null) | 1186 | CLOSE |
| 1187 | NEW | 1187 | SR | WORKORDER | 1240 | 1240 | CLOSE | (null) | (null) |
| 1187 | NEW | 1187 | SR | SR | 1188 | (null) | (null) | 1188 | NEW |
| 1190 | NEW | 1190 | SR | SR | 1191 | (null) | (null) | 1191 | CLOSE |
| 1192 | NEW | 1192 | SR | WORKORDER | 1241 | 1241 | CLOSE | (null) | (null) |
+----------+--------+-----------+-------+-----------------+---------------+--------+--------+--------+--------+
所以
从这个集合中,我只想在结果集中看到票证1185,1190,1192。 它应该看起来像:
+----------+--------+-----------+-------+-----------------+---------------+--------+--------+--------+--------+
| ticketid | status | RECORDKEY | class | RELATEDRECCLASS | relatedreckey | id | status | id | status |
+----------+--------+-----------+-------+-----------------+---------------+--------+--------+--------+--------+
| 1185 | NEW | 1185 | SR | WORKORDER | 1239 | 1239 | CLOSE | (null) | (null) |
| 1185 | NEW | 1185 | SR | SR | 1186 | (null) | (null) | 1186 | CLOSE |
| 1190 | NEW | 1190 | SR | SR | 1191 | (null) | (null) | 1191 | CLOSE |
| 1192 | NEW | 1192 | SR | WORKORDER | 1241 | 1241 | CLOSE | (null) | (null) |
+----------+--------+-----------+-------+-----------------+---------------+--------+--------+--------+--------+
我尝试过类似的事情:
select ticket.ticketid, rr.relatedrecordkey, rr.relatedrecordclass, c1.id, c1.status, c2.id, c2.status
from ticket
inner join relatedrecord rr on rr.recordkey = ticket.ticketid and rr.class = ticket.class
where not exists (
select 1 from child1 c1
where c1.id = rr.relatedreckey and c1.class = rr.relatedrecclass
and c1.status <> 'CLOSE'
) and not exists (
select 1 from child2 c2
where c2.id = rr.relatedreckey and c2.class = rr.relatedrecclass
and c2.status <> 'CLOSE'
)
这导致ticket2的两行和来自ticket 3的一行(因为它关闭了一个子) 我有点困惑如何正确解决这个问题。
答案 0 :(得分:0)
仍然不能完全确定你想要的结果,但我认为这可能是它。
select *
from
(
select t.ticketid
from ticket t
join relatedrecord rr on rr.recordkey = t.ticketid and rr.class = t.class
join child1 c1 on c1.id = rr.relatedreckey
and c1.class = rr.relatedrecclass
and c1.status = 'CLOSE'
group by t.ticketid
having MAX(c1.status) = MIN(c1.status)
UNION ALL
select t.ticketid
from ticket t
join relatedrecord rr on rr.recordkey = t.ticketid and rr.class = t.class
join child2 c2 on c2.id = rr.relatedreckey
and c2.class = rr.relatedrecclass
and c2.status = 'CLOSE'
group by t.ticketid
having MAX(c2.status) = MIN(c2.status)
) x
group by ticketid
答案 1 :(得分:0)
我找到了合适的答案。
请参阅此sqlfiddle:http://sqlfiddle.com/#!3/febde/24
SELECT ticket.ticketid,
count(rr.relatedreckey) AS children,
count(wo.id) AS wo,
count(tt.id) AS sr
FROM ticket
INNER JOIN relatedrecord rr ON rr.recordkey = ticket.ticketid
AND rr.class = ticket.class
LEFT JOIN child1 wo ON wo.id = rr.relatedreckey
AND wo.class = rr.relatedrecclass
AND wo.status = 'CLOSE'
LEFT JOIN child2 tt ON tt.id = rr.relatedreckey
AND tt.class = rr.relatedrecclass
AND tt.status = 'CLOSE'
GROUP BY ticket.ticketid
HAVING (count(wo.id) + count(tt.id)) = count(rr.relatedreckey);
基本上它为每个相关类型计算已关闭记录的数量,并将其与相关记录的总数进行比较。如果它们相同,则必须关闭所有相关记录。