带表达式的SSRS困境

Question

我有一个SSRS数据集，如下所示：

使用UNION ALL生成数据集行，彼此独立。 我需要在报告中按原样显示这些行，但是我需要添加一个额外的行来计算Total Won / Total Lost，所以结果应如下所示：

这只是样本，因为我有更多的列（每月1个），整个事情按产品细分，所以如果我有10个不同的产品，我将有10个不同的tablix表。

基本上我需要以某种方式创建一个表达式，该表达式只计算3个Tablix中2行的值（基于Status列的值），并考虑到某些值可以为零。

这里是查询（为了更好地理解我简化了一下）：

select * from
(
select 'Created' as 'State', fo.groupidname, fo.businessidname ' Business', fo.opportunityid
from FilteredOpportunity fo
where fo.regionidname = 'Americas Region'
and fo.createdon >= dateadd(year, -1, getdate())
and fo.regionalfeeincome >= 250000
) created
pivot
(
count(created.opportunityid)
for created.groupidname in ([Boston], [Chicago], [Colombia], [Group D.C.], [Houston], [Los Angeles], [New York], [San Francisco], [Seattle], [Toronto])
) pivCreated
union all
select * from
(
select 'Won' as 'State', fo.groupidname, fo.businessidname ' Business', fo.opportunityid
from FilteredOpportunity fo
where regionidname = 'Americas Region'
and fo.actualclosedate >= dateadd(year, -1, getdate())
and regionalfeeincome >= 250000
and fo.jna is not null 
) won
pivot
(
count(won.opportunityid)
for won.groupidname in ([Boston], [Chicago], [Colombia], [Group D.C.], [Houston], [Los Angeles], [New York], [San Francisco], [Seattle], [Toronto])
) pivWon
union all
select * from
(
select 'Lost' as 'State', fo.groupidname, fo.businessidname ' Business', fo.opportunityid
from FilteredOpportunity fo
where fo.regionidname = 'Americas Region'
and fo.actualclosedate >= dateadd(year, -1, getdate())
and fo.regionalfeeincome >= 250000
and fo.sys_phasename <> 'Pre-Bid'
) lost
pivot
(
count(lost.opportunityid)
for lost.groupidname in ([Boston], [Chicago], [Colombia], [Group D.C.], [Houston], [Los Angeles], [New York], [San Francisco], [Seattle], [Toronto])
) pivLost

TIA -TS。

Answer 1

我无法对此进行全面测试，因为我没有时间构建样本数据，但它应该可以工作......

如果您将此作为报告的数据集查询，那么您应该能够添加一个简单的矩阵，其中State的行组和City的列组

/*
You can optimise this a bit but I've kept it fairly procedural so it's easier to follow
*/
-- First do your 3 basic queries but this time we don't pivot and we dump the results into temp tables
-- I've also removed columns that don't appear to be used based on your sample output and remaned a column to City to make it easier to read
select 'Total Created' as 'State', fo.groupidname as City, COUNT(*) AS Cnt
INTO #Created
    from FilteredOpportunity fo
    where fo.regionidname = 'Americas Region'
        and fo.createdon >= dateadd(year, -1, getdate())
        and fo.regionalfeeincome >= 250000

select 'Total Won' as 'State', fo.groupidname as City, COUNT(*) AS Cnt
INTO #Won
    from FilteredOpportunity fo
    where fo.regionidname = 'Americas Region'
        and fo.createdon >= dateadd(year, -1, getdate())
        and fo.regionalfeeincome >= 250000
        and fo.jna is not null 

select 'Total Lost' as 'State', fo.groupidname as City, COUNT(*) AS Cnt
INTO #Lost
    from FilteredOpportunity fo
    where fo.regionidname = 'Americas Region'
        and fo.createdon >= dateadd(year, -1, getdate())
        and fo.regionalfeeincome >= 250000
        and fo.sys_phasename <> 'Pre-Bid'

-- Now we calculate your Ratio
SELECT
    'Win Ratio' as 'State' 
    , coalesce(w.City, l.City) as City -- use coalesce in case 1 side of join is null
    , CASE WHEN ISNULL(l.cnt,0) = 0 THEN 0 ELSE w.Cnt/l.Cnt END as Cnt -- if lost is null or 0 return 0 else calc ratio
into #Ratio
 FROM #Won w
    full join #Lost l on w.City = l.City

-- finaly, get the results and add a sort to help the report row sorting.
SELECT 1 as Sort, * FROM #Created
UNION SELECT 2, * FROM #Won
UNION SELECT 3, * FROM #Lost
UNION SELECT 4, * FROM #Ratio