0

使用 11gR2，windows 7 客户端机器。我需要更新表missing_volume（如下），我需要在其中计算estimated_missing 列。本月的estimated_missing 列的计算需要上个月的数字（如下代码中所述）。我想要像第一个表一样的输出。请注意，记录从一月份开始，因此无法计算一月份的estimated_missing，但对于其余月份，只需更改“yr”和“mnth”即可完成（在代码中最后注释）。

yr      mnth        location    volume      actual_missing    expected_missing  estimated_missing

2013    January     loc1        48037       24                57                    
2013    February    loc1        47960       3660              53                24
2013    March       loc1        55007       78                57                28
2013    April       loc1        54345       72                58                77

编码：

UPDATE missing_volume g
    SET
        g.estimated_missing = 
        (
            SELECT

                CASE
                    -- If the previous month actual_missing number is less than 2 times than that of the previous month expected_missing number,
                    -- then for current month: 
                    -- estimated_missing = previous_month(actual_missing/volume)*current_month(volume)
                    -- 
                    -- Else, for current month:
                    -- estimated_missing = previous_month(estimated_missing/volume)*current_month(volume)

                    WHEN (NVL(p.actual_missing, 0)/NVL(p.expected_missing, 1)) <= 2.0
                        THEN ROUND(NVL(p.actual_missing, 0)*( NVL(g.volume, 0)/NVL(p.volume, 1) ), 0)
                    ELSE
                        ROUND(NVL(p.estimated_missing, 0)*( NVL(g.volume, 0)/NVL(p.volume, 1) ), 0)
                END                     

            FROM 
                (
                    SELECT
                        yr,
                        mnth,
                        loc_cd,
                        volume,
                        actual_missing,
                        expected_missing,   
                        estimated_missing
                    FROM missing_volume
                    WHERE yr = TRIM(TO_CHAR(ADD_MONTHS(SYSDATE, -2), 'YYYY'))   -- For February it is -4, for March it is -3, for April it is -2
                    AND UPPER(mnth) = UPPER(TRIM(TO_CHAR(ADD_MONTHS(SYSDATE, -2), 'Month')))  -- For February it is -4, for March it is -3, for April it is -2
                )p
            WHERE g.yr = TRIM(TO_CHAR(ADD_MONTHS(SYSDATE, -1), 'YYYY'))   -- For February it is -3, for March it is -2, for April it is -1
            AND UPPER(g.mnth) = UPPER(TRIM(TO_CHAR(ADD_MONTHS(SYSDATE, -1), 'Month')))  -- For February it is -3, for March it is -2, for April it is -1
            AND g.volume IS NOT NULL
            AND TRIM(p.location) = TRIM(g.location)                     
        );

当我为每个月运行代码时，代码确实为“estimated_missing”计算了正确的数字，但问题是在更新当前月份时，它也会删除上个月的记录。例如，如下所示，在我更新四月后，该列只有四月的记录，上个月的记录消失了，同样更新三月删除了二月等。我不明白为什么会这样！这是我得到的输出：

yr      mnth        location    volume      actual_missing    expected_missing  estimated_missing

2013    January     loc1        48037       24                57                    
2013    February    loc1        47960       3660              53                
2013    March       loc1        55007       78                57                
2013    April       loc1        54345       72                58                77

如果有人能解释为什么会发生这种情况（我的意思是代码中的缺陷在哪里）以及如何获得所需的输出（第一个表），将不胜感激。

谢谢

4

1 回答 1

0

您需要为语句指定 where 子句。目前 WHERE 子句在大括号内。你的陈述基本上是：

UPDATE missing_volume g
SET
    g.estimated_missing = 
    ( [...]
    ); -- no where clause here

你需要的是这样的：

UPDATE missing_volume g
SET
    g.estimated_missing = 
    ( [...]
    )
WHERE g.yr = TRIM(TO_CHAR(ADD_MONTHS(SYSDATE, -1), 'YYYY'));

自己没有测试过，但这应该只更新当月的行。

于 2013-07-04T16:06:23.453 回答