Question

我现在正在维护一个很棒的应用程序，其他人写了从一些大型政府遗留系统中挖掘出一些数据。基本上我需要一个查询结果来填充一个gridview，它从Tech Order获取每个零件号，并计算Fedlog表中匹配的零件号。如果没有找到，那么请查看“商业”表。现有查询目前仅在Fedlog表中查找，内容如下：

select p.*,
 (select case when count(*) > 0 then 'Y' else 'N' end as SL
  from tbl_fedlog where [Part Number] = p.[Part Number]) as SL
from tbl_pcms p 
where  p.[Tech Order] = '0B-E0C-9' order by p.Figure, p.[Index], p.Indenture

当'N'时，我必须查看商业表格。我可以就最好的方法提出一些建议吗？

Answer 1

“N”情景的结果应该是什么 - 只是另一个“Y”/“N”答案？如果是这样，您应该能够简单地将“N”表达式替换为针对“商业”表的另一个标量查询。

Answer 2

这是Fedlog查询和商业查询之间的UNION ALL，它检查每个查询是否存在：

SELECT p.*
, SL = 'Y'
, part_count = (SELECT COUNT(*) FROM tbl_fedlog WHERE [Part Number] = p.[Part Number])
from tbl_pcms p 
where  p.[Tech Order] = '0B-E0C-9'
AND EXISTS(SELECT 1 FROM tbl_fedlog WHERE [Part Number] = p.[Part Number])

UNION ALL

SELECT p.*
, SL = 'Y'
, part_count = (SELECT COUNT(*) FROM tbl_commercial WHERE [Part Number] = p.[Part Number])
from tbl_pcms p 
where  p.[Tech Order] = '0B-E0C-9'
AND EXISTS(SELECT 1 FROM tbl_commercial WHERE [Part Number] = p.[Part Number])
AND NOT EXISTS(SELECT 1 FROM tbl_fedlog WHERE [Part Number] = p.[Part Number])

我需要更多信息来解决其他答案中提到的'N'场景。

Answer 3

如果逻辑是在'Y'或Fedlog表格中存在该部分时显示Commercial，否则'N'，则可以尝试分组＆amp;分别聚合这些表并将（聚合的）聚合结果集加入tbl_pcms，如下所示：

SELECT
  p.*,
  CASE WHEN COALESCE(f.PartCount, c.PartCount) IS NULL THEN 'N' ELSE 'Y' END AS SL
FROM tbl_pcms p

  LEFT JOIN (
    SELECT
      [Part Number],
      COUNT(*) AS PartCount
    FROM tbl_fedlog
    GROUP BY [Part Number]
  ) f ON p.[Part Number] = f.[Part Number]

  LEFT JOIN (
    SELECT
      [Part Number],
      COUNT(*) AS PartCount
    FROM tbl_commercial
    GROUP BY [Part Number]
  ) c ON p.[Part Number] = c.[Part Number]

WHERE p.[Tech Order] = '0B-E0C-9'
ORDER BY
  p.Figure,
  p.[Index],
  p.Indenture

是否有可能在一个表中“不在/存在”然后在另一个表中？

3 个答案: