Question

这一直很令人费解，希望有人可以赐教。我有一个存储过程，当我从SSMS运行它时返回预期（至少对我来说）结果，并且当我从SQL作业运行它时返回意外结果。服务器正在运行SQL Server 2005，不确定sp版本。如果它有帮助我可以得到它。

当我从SSMS运行下面的代码时，只有一条记录，Mon，Mar 05：，包含InsufficientSampleSize的'True'，正如我所料。

当我从SQL作业运行相同的代码时，所有记录都包含InsufficientSampleSize的“True”，这是意外的。

我是如何评估d.ScoreCount值的？也许这是一个像ANSI_NULLS这样的数据库设置，当我使用Job运行程序时会有所不同？我的数据目前不包含任何空值，所以我认为它不是ANSI_NULLS。

提前致谢。我尝试过多次重写这个逻辑，但我仍然遇到这个问题。

create table #tmpScoreData (EventDate datetime, Score int)

insert into #tmpScoreData values('20120302', 4)
insert into #tmpScoreData values('20120302', 7)
insert into #tmpScoreData values('20120302', 7)
insert into #tmpScoreData values('20120302', 2)
insert into #tmpScoreData values('20120302', 7)

insert into #tmpScoreData values('20120303', 9)
insert into #tmpScoreData values('20120303', 2)
insert into #tmpScoreData values('20120303', 4)
insert into #tmpScoreData values('20120303', 2)
insert into #tmpScoreData values('20120303', 5)

insert into #tmpScoreData values('20120304', 8)
insert into #tmpScoreData values('20120304', 8)
insert into #tmpScoreData values('20120304', 7)
insert into #tmpScoreData values('20120304', 2)
insert into #tmpScoreData values('20120304', 5)

insert into #tmpScoreData values('20120305', 7)
insert into #tmpScoreData values('20120305', 6)
insert into #tmpScoreData values('20120305', 2)

insert into #tmpScoreData values('20120306', 7)
insert into #tmpScoreData values('20120306', 2)
insert into #tmpScoreData values('20120306', 4)
insert into #tmpScoreData values('20120306', 2)
insert into #tmpScoreData values('20120306', 5)

insert into #tmpScoreData values('20120307', 5)
insert into #tmpScoreData values('20120307', 2)
insert into #tmpScoreData values('20120307', 8)
insert into #tmpScoreData values('20120307', 7)
insert into #tmpScoreData values('20120307', 5)

insert into #tmpScoreData values('20120308', 9)
insert into #tmpScoreData values('20120308', 4)
insert into #tmpScoreData values('20120308', 8)
insert into #tmpScoreData values('20120308', 3)
insert into #tmpScoreData values('20120308', 5)

declare @weekEnding datetime
set @weekEnding = '20120308'

select EventDate = cd.WeekDayAbbr + ', ' + substring(convert(varchar(10), cd.CalendarDate, 107), 1, 6) + ':',
    AvgScore = d.Score,
    InsufficientSampleSize = case when isnull(d.ScoreCount, 0) < 5 then 'True' end,
    WeekEnding = @weekEnding
from (select 1 as ID, 'Fri' as WeekDayAbbr, @weekEnding - 6 as CalendarDate
    union
    select 2, 'Sat', @weekEnding - 5
    union
    select 3, 'Sun', @weekEnding - 4
    union
    select 4, 'Mon', @weekEnding - 3
    union
    select 5, 'Tue', @weekEnding - 2
    union
    select 6, 'Wed', @weekEnding - 1
    union
    select 7, 'Thu', @weekEnding) as cd
left join (select EventDate,
                avg(convert(decimal(9,4), Score)) as Score,
                count(Score) as ScoreCount
           from #tmpScoreData
          where EventDate <= @weekEnding
            and EventDate > @weekEnding - 7
          group by EventDate) as d on cd.CalendarDate = d.EventDate
order by cd.ID

drop table #tmpScoreData

我将日期更新为不太模糊的值。在我的实际存储过程中，日期在一个exising表中作为datetime值。所以，这不是我的问题的原因。感谢。

Answer 1

您的日期可能会有不同的解释。 '3/8/2012'是指8月3日还是3月8日？

为避免歧义，请不要只使用字符串和隐式转换。而是使用字符串和显式转换......

CONVERT(DATETIME, '3/8/2012', 103)  -- Or whichever format is relevant

有关您需要的格式，请参阅CAST and CONVERT (Transact-SQL)。

SQL Server 2005 SSMS和SQL作业之间的查询结果不一致

1 个答案: