Question

我正在尝试将公式中的某些输入字符串替换为零值。我能够为某些字符串执行此操作，但无法对下面#mathtemp表中列出的所有字符串执行此操作。

我有输入名称，我有公式，我需要做的只是将没有数字（值）的字符串重新归零。替换语句中有一些错误，我无法弄清楚。

我尝试使用ISNUMERIC来完成此操作，但这并未提供所需的输出。如果有任何新想法可以解决以下问题，请告诉我。

BEGIN
DECLARE @INPUTCOUNT AS INTEGER
DECLARE @formulacount as integer
DECLARE @in as int
DECLARE @ia as int
DECLARE @OUTPUTFORMULATrade as nvarchar(160)
set @ia = 1
set @in = 1



CREATE TABLE #STATICFILTER
(
IDNUM INTEGER IDENTITY(1,1),
STATICNAME Varchar(160),
)

INSERT INTO #STATICFILTER (STATICNAME)
VALUES ('Import - Consumption customs value(2266)') ,('Import - Consumption customs value(1540)') ,('Import - Consumption customs value(1541)')


SET @INPUTCOUNT = (select count(*)  from #STATICFILTER)

CREATE TABLE #MathTemp1
(
IDNUM INTEGER IDENTITY(1,1),
YEARMONTH VARCHAR(256),
OUTPUTFORMULA VARCHAR(256),
Timedimensiondate Date
)

INSERT INTO #MathTemp1 (YEARMONTH,OUTPUTFORMULA,Timedimensiondate)
VALUES ('CV(N2)  1989: 1','2641.000 + Import - Consumption customs value(1540) + Import - Consumption customs value(1541)','1989-01-01')
,('CV(N2)  1989: 10','54407.000 + Import - Consumption customs value(1540) + 63906.000','1989-10-01')
,('CV(N2)  1990: 11','Import - Consumption customs value(2266) + Import - Consumption customs value(1540) + 53088.000','1990-11-01')
,('CV(N2)  1994: 5','32852.000 + Import - Consumption customs value(1540) + Import - Consumption customs value(1541)','1994-05-01')
SET @formulacount = (select count(*)  from #MathTemp1)

while (@ia < @formulacount)
begin
WHILE (@in < @INPUTCOUNT)
BEGIN
SET @OUTPUTFORMULATrade = (Select Replace ((SELECT REPLACE(OUTPUTFORMULA,(select STATICNAME from #STATICFILTER where IDNUM = @in),0)FROM #MathTemp1 WHERE IDNUM = @ia),(select STATICNAME from #STATICFILTER where IDNUM = @in+1),0))
SET @in = @in + 1
END
SET @ia = @ia + 1
SET @in  =1
Select @OUTPUTFORMULATrade AS New 
END
drop table #MathTemp1
drop table #STATICFILTER

END

请注意输入在#staticfilter表中以随机格式排列，同样#mathtemp表也可能包含输入公式的任意组合。

我想要的输出是将任何给定公式中的字符串替换为零值。请查看以下

2641.000 + 0 + 0
54407.000 + 0 + 63906.000
0 + 0 + 53088.000

Answer 1

This takes a number of somewhat advanced techniques combined to do this. The first problem is you have delimited data. This violates 1NF when you cram multiple values into a single cell. The second piece of the puzzle is how to PIVOT this data into a dynamic number of columns. Most people around SO prefer to use a dynamic PIVOT. I prefer to use a dynamic cross tab instead. I find the syntax less obtuse and it is even a little bit more performant than a dynamic cross tab.

You can read about the splitter I typically use here. http://www.sqlservercentral.com/articles/Tally+Table/72993/ The main advantage this splitter offers that most others don't is that it returns the row number of the item within the list of values. This is incredibly useful for this type of situation. If you really want to dive into the splitter world here are several other excellent options. http://sqlperformance.com/2012/07/t-sql-queries/split-strings

You can read more about Dynamic cross tabs here. http://www.sqlservercentral.com/articles/Crosstab/65048/

I don't really understand what the #STATICFILTER table has to do with this so I just ignored it.

Make sure you understand this code before you implement it. The referenced articles go into great detail about these techniques.

if OBJECT_ID('tempdb..#MathTemp1') is not null
    drop table #MathTemp1

CREATE TABLE #MathTemp1
(
    IDNUM INTEGER IDENTITY(1,1),
    YEARMONTH VARCHAR(256),
    OutputFormula VARCHAR(256),
    Timedimensiondate Date
)

INSERT INTO #MathTemp1 (YEARMONTH,OUTPUTFORMULA,Timedimensiondate)
VALUES ('CV(N2)  1989: 1','2641.000 + Import - Consumption customs value(1540) + Import - Consumption customs value(1541)','1989-01-01')
,('CV(N2)  1989: 10','54407.000 + Import - Consumption customs value(1540) + 63906.000','1989-10-01')
,('CV(N2)  1990: 11','Import - Consumption customs value(2266) + Import - Consumption customs value(1540) + 53088.000','1990-11-01')
,('CV(N2)  1994: 5','32852.000 + Import - Consumption customs value(1540) + Import - Consumption customs value(1541)','1994-05-01')

declare @StaticPortion nvarchar(2000) = 
    'with OrderedResults as
    (   
        select mt.IDNUM
            , mt.OutputFormula
            , mt.Timedimensiondate
            , mt.YEARMONTH
            , x.ItemNumber
            , LTRIM(RTRIM(x.Item)) as Item
        from #MathTemp1 mt
        cross apply dbo.DelimitedSplit8K(mt.OutputFormula, ''+'') x
    )
    Select IDNUM';

declare @DynamicPortion nvarchar(max) = '';
declare @FinalStaticPortion nvarchar(2000) = ' from OrderedResults Group by IDNUM order by IDNUM';  

with E1(N) AS (select 1 from (values (1),(1),(1),(1),(1),(1),(1),(1),(1),(1))dt(n)),
E2(N) AS (SELECT 1 FROM E1 a, E1 b), --10E+2 or 100 rows
E4(N) AS (SELECT 1 FROM E2 a, E2 b), --10E+4 or 10,000 rows max
cteTally(N) AS 
(
    SELECT  ROW_NUMBER() OVER (ORDER BY (SELECT NULL)) FROM E4
)

select @DynamicPortion = @DynamicPortion + 
    ', MAX(Case when ItemNumber = ' + CAST(N as varchar(6)) + 'then case when ISNUMERIC(Item) = 1 then convert(numeric(9,3), ltrim(rtrim(Item))) else 0 end end) as Value' + CAST(N as varchar(6)) + CHAR(10)
from cteTally t
where t.N <= 
(
    select MAX(LEN(OutputFormula) - LEN(replace(OutputFormula, '+', ''))) + 1
    from #MathTemp1
)


declare @SqlToExecute nvarchar(max) = @StaticPortion + @DynamicPortion + @FinalStaticPortion;

--select @SqlToExecute
exec sp_executesql @SqlToExecute

在SQL

1 个答案: