1

以下代码可以正常工作：

select donem, mekankodu, count(yayin_kodu) yc, 
       SUM(CASE WHEN iade =0 THEN 1 ELSE 0 END) yys
from 
(    select donem,
            bayi_adi,
            bayi_kodu,
            x.mekankodu,
            mekan_adi,
            mekan_tipi,
            yayin_kodu,
            yayin_adi,
            sum(x.b2b_dagitim + x.b2b_transfer) sevk,
            sum(x.b2b_iade) iade,
            sum(x.b2b_dagitim + x.b2b_transfer - x.b2b_iade) satis
    from mps_view2 x
    where x.donem = '200910'
    and x.yayin_kodu in (93,95,98,99,103,174,307,309,311,489,491,495,533,534,538,605,641,642,650,2539,
                         2697,4560,5049,5772,5950,5980,7318,7440,8086,8524,11161,12707,12708,12709,14376,
                         15107,15309,15633)
    and x.mekankodu in (31851,38569,7123,7403,7481)
    group by donem, bayi_adi, bayi_kodu, mekankodu, mekan_adi, mekan_tipi, yayin_kodu, yayin_adi
    order by donem, bayi_kodu, mekan_adi, yayin_adi
)
group by donem, mekankodu

当我想包含yc/yss在我的结果表中时，我将SELECT部分安排如下：

select donem, mekankodu, count(yayin_kodu) yc, 
       SUM(CASE WHEN iade =0 THEN 1 ELSE 0 END) yys,
       yc/yys yo

但它给了

ORA-00904 错误：YYS：标识符无效。

我该怎么办？

4

4 回答 4

6

使用 with 语句：

with subquery_name as
(
    select
       donem,
       mekankodu,
       count(yayin_kodu) as yc, 
       SUM(CASE WHEN iade =0 THEN 1 ELSE 0 END) as yys
    from
        ...
)
select
    donem,
    mekankodu,
    yc,
    yys,
    yc/yys as yo
from
    subquery_name

这是 PL/SQL 的一个非常有用的特性。您正在尝试访问同一查询中生成的列，这是不可能的。with 语句将允许您像在此处所做的那样基于另一个计算列来计算新列。

于 2010-01-28T13:20:14.610 回答
3

您不能在同一选择中引用其他列。使用子选择：

select donem, mekankodu, yc, yys, yc/yys yo
from
(
    select donem, mekankodu, count(yayin_kodu) yc,
        SUM(CASE WHEN iade =0 THEN 1 ELSE 0 END) yys,
        yc/yys yo
    from ...
) Sub
于 2010-01-28T13:20:36.033 回答
2

您不能访问 SQL 中相邻创建的列，但是可以访问数据目录：

select donem, mekankodu, count(yayin_kodu) yc, 
       SUM(CASE WHEN iade =0 THEN 1 ELSE 0 END) yys,
       count(yayin_kodu)/SUM(CASE WHEN iade =0 THEN 1 ELSE 0 END) "yc/yys"
from 
(    select donem,
            bayi_adi,
            bayi_kodu,
            x.mekankodu,
            mekan_adi,
            mekan_tipi,
            yayin_kodu,
            yayin_adi,
            sum(x.b2b_dagitim + x.b2b_transfer) sevk,
            sum(x.b2b_iade) iade,
            sum(x.b2b_dagitim + x.b2b_transfer - x.b2b_iade) satis
    from mps_view2 x
    where x.donem = '200910'
    and x.yayin_kodu in (93,95,98,99,103,174,307,309,311,489,491,495,533,534,538,605,641,642,650,2539,
                         2697,4560,5049,5772,5950,5980,7318,7440,8086,8524,11161,12707,12708,12709,14376,
                         15107,15309,15633)
    and x.mekankodu in (31851,38569,7123,7403,7481)
    group by donem, bayi_adi, bayi_kodu, mekankodu, mekan_adi, mekan_tipi, yayin_kodu, yayin_adi
    order by donem, bayi_kodu, mekan_adi, yayin_adi
)
group by donem, mekankodu

注意：这并不像看起来那样效率低下……SQL 足够聪明，可以在大多数情况下进行一次计算并使用两次结果。通常情况下，相信优化器。

于 2010-01-28T13:20:13.127 回答
-1

你忘了“as”吗？

select donem, mekankodu, count(yayin_kodu) yc, 
   SUM(CASE WHEN iade =0 THEN 1 ELSE 0 END) AS yys
于 2010-01-28T13:18:01.433 回答