Question

嗨，我试图用两个以上的表创建一个数据表，但是我的返回数据集有一个空值。我想使它返回0并同时使用ISNULL和COALESCE，但是自从我使用Pivot以来，这两个表都没有对我起作用操作。以下是我的SQL查询。

SELECT MetricHeader,[16],[17],[18],[22],[23],[24],[26],[27],[28],[29],[30]
FROM (
SELECT 'P1' as MetricHeader,co.OptionId as CrossbreakHeader,
ISNULL(CONVERT(decimal,k.Response),0) as KpiResponse
FROM
( 
SELECT PanelistId,ProjectId FROM zcachetable_Edm_Sampling_1_5  
INTERSECT
SELECT PanelistId,ProjectId FROM zcachetable_Edm_Sampling_1_5  
INTERSECT
SELECT PanelistId,ProjectId FROM zcachetable_Edm_Responses_1_17_51 as d
WHERE d.ResponseCode IN (42,43,44,45,46,47,48,49,50,51)  
INTERSECT
SELECT PanelistId,ProjectId FROM zcachetable_Edm_Responses_1_17_54 as d
WHERE d.ResponseCode IN (16,17,18,22,23,24,26,27,28,29,30)
INTERSECT
SELECT PanelistId,ProjectId FROM zcachetable_Edm_Responses_1_17_55 as d
WHERE d.ResponseCode IN (19,20,21,31,32,33,34,35)
) f
INNER JOIN zcachetable_Edm_Responses_1_17_54 cc ON cc.PanelistId = f.PanelistId AND f.ProjectId=cc.ProjectId
INNER JOIN n1 co ON ((cc.ResponseCode IS NOT NULL AND cc.ResponseCode = co.OptionId))
INNER JOIN zcachetable_Edm_Responses_1_17_55_KPIScore k ON cc.PanelistId=k.PanelistId AND cc.ProjectId=k.ProjectId
) sq
PIVOT
(
AVG(KpiResponse)
FOR sq.CrossbreakHeader IN ([16],[17],[18],[22],[23],[24],[26],[27],[28],[29],[30])
) AS pt

我已经附加了结果集的图片，并且它的列值30为空值。我想将其设置为0，并且尝试了Internet上的每个选项都找不到解决方案。我将非常感谢你们可以建议我这样做。

Answer 1

您可以使用ISNULL()或COALESCE()处理NULL值

示例，您的查询

ISNULL ( [30], 0)

Answer 2

进行透视时，全神贯注于此非常容易，但实际上，这是您需要使用COALESE或ISNULL的最终select子句，如下所示：

SELECT
    MetricHeader
  , COALESCE([16], 0) AS [16]
  , COALESCE([17], 0) AS [17]
  , COALESCE([18], 0) AS [18]
  , COALESCE([22], 0) AS [22]
  , COALESCE([23], 0) AS [23]
  , COALESCE([24], 0) AS [24]
  , COALESCE([26], 0) AS [26]
  , COALESCE([27], 0) AS [27]
  , COALESCE([28], 0) AS [28]
  , COALESCE([29], 0) AS [29]
  , COALESCE([30], 0) AS [30]
FROM (
    SELECT
        'P1'                                    AS metricheader
      , co.OptionId                             AS crossbreakheader
      , ISNULL(CONVERT(decimal, k.Response), 0) AS kpiresponse
    FROM (
        SELECT
            PanelistId
          , ProjectId
        FROM zcachetable_Edm_Sampling_1_5

        INTERSECT

        SELECT
            PanelistId
          , ProjectId
        FROM zcachetable_Edm_Sampling_1_5

        INTERSECT

        SELECT
            PanelistId
          , ProjectId
        FROM zcachetable_Edm_Responses_1_17_51 AS d
        WHERE d.ResponseCode IN (42, 43, 44, 45, 46, 47, 48, 49, 50, 51)

        INTERSECT

        SELECT
            PanelistId
          , ProjectId
        FROM zcachetable_Edm_Responses_1_17_54 AS d
        WHERE d.ResponseCode IN (16, 17, 18, 22, 23, 24, 26, 27, 28, 29, 30)

        INTERSECT

        SELECT
            PanelistId
          , ProjectId
        FROM zcachetable_Edm_Responses_1_17_55 AS d
        WHERE d.ResponseCode IN (19, 20, 21, 31, 32, 33, 34, 35)
    ) f
    INNER JOIN zcachetable_Edm_Responses_1_17_54 cc ON cc.PanelistId = f.PanelistId
        AND f.ProjectId = cc.ProjectId
    INNER JOIN n1 co ON ((cc.ResponseCode IS NOT NULL
        AND cc.ResponseCode = co.OptionId))
    INNER JOIN zcachetable_Edm_Responses_1_17_55_KPIScore k ON cc.PanelistId = k.PanelistId
        AND cc.ProjectId = k.ProjectId
    ) sq
PIVOT
    (
    AVG(KpiResponse)
    FOR sq.CrossbreakHeader IN ([16], [17], [18], [22], [23], [24], [26], [27], [28], [29], [30])
    ) AS pt

Answer 3

@Squirrel为您提供了解决此问题的技巧。 ISNULL已发生，然后再进行透视。对于CrossbreakHeader = '30'，没有行，仅此而已。

–伊万·斯塔诺斯汀

create table #null ( i int, name varchar(10))

insert #null values
(1,'One'),(2,'Two'),(3,'Three'),(4,'Four'),(null,'Five')

select one, five, six from ( -- Here, NULL for 'six'
 select name, ISNULL(i, 0) i from #null
) n
pivot
(
 sum(i) for name in (one, five, six)  -- I cant insert 'six'
)p

ISNULL或COALESCE不适用于联接和枢轴操作

3 个答案: