Question

这是我一直在使用的查询：

select serial_number, DAQ.qtag_no, min(TOH.start_time), order_number, DAQ.creation_time, qtag_status, ar_code, pro_foundin, pro_category, root_cause, remark from UNIT U 
left JOIN WORK_ORDER WO ON WO.order_key = U.order_key
left join TRACKED_OBJECT_HISTORY TOH on TOH.tobj_key = U.unit_key
WHERE U.creation_time > '01/01/2012' AND U.creation_time < '07/20/2012'
AND order_number NOT LIKE '[R]%' 
group by serial_number, qtag_no, order_number, DAQ.creation_time, qtag_status, ar_code, pro_foundin, pro_category, root_cause, remark
order by serial_number

现在我得到3280个结果。在此设置中，有不同的站，例如＆＃34;组件，＆＃34; ＆＃34;诊断，＆＃34;我的min（TOH.start_time）列的目标是返回装配站的第一个开始时间，但目前它在任何一个站点返回第一个开始时间。但是，如果我添加另一个WHERE子句来指定工作站（TOH.op_name =＆＃39; Assembly＆＃39;），它会限制结果数量（低至2700）。我想保留3280结果，而不是扫描到装配站的单元，为min（TOH.start_time）列返回NULL。我尝试使用case函数，但这要求我在group by子句中包含TOH.op_name，这是我不想要的。谢谢！

Answer 1

你应该能够将你的条件添加到我认为的左连接中，如下所示：

select  serial_number, 
    DAQ.qtag_no, 
    min(TOH.start_time), 
    order_number, 
    DAQ.creation_time, 
    qtag_status, 
    ar_code, 
    pro_foundin, 
    pro_category, 
    root_cause, 
    remark 
from UNIT U 
left 
JOIN    WORK_ORDER WO 
    ON  WO.order_key = U.order_key
left 
join    TRACKED_OBJECT_HISTORY TOH 
    on  TOH.tobj_key = U.unit_key
    and TOH.op_name = 'Assembly'
WHERE   U.creation_time > '01/01/2012' AND U.creation_time < '07/20/2012'
AND     order_number NOT LIKE '[R]%' 
group by serial_number, qtag_no, order_number, DAQ.creation_time, qtag_status, ar_code, pro_foundin, pro_category, root_cause, remark
order by serial_number

如何在不更改结果数量的情况下获取聚合结果SQL

1 个答案: