Question

我有使用主要版本和次要版本进行版本控制的文件现在我想要做的就是获取所有文件的最高主要版本（我已经完成了）。然后检查次要版本以查看它是否包含0.如果它确实丢弃了整套版本。所以举个例子。我的第一个查询返回：

FileA Ver 1.0, 1.1, 1.2, 1.3 
FileB Ver 2.1, 2.2, 2.3, 2.4 
FileC Ver 5.1, 5.2, 5.3.

所以在所有11行/记录中。现在我的第二个查询应该采用该结果并抛弃所有FileA版本，因为其中一个版本的0为次要版本。所以第二个查询应该返回：

FileB Ver 2.1, 2.2, 2.3, 2.4 
FileC Ver 5.1, 5.2, 5.3.

共7行/记录。有人可以帮我解决这个问题吗？如果有帮助，我正在使用SQL Server 2008。

latestFileMajorVersion(fileId, majRev)
as
(
    --Gets files with highest major version
    select distinct v_fileid, max(v_majrev)
    from files
    group by v_fileid
),
latestFileVersions(fileId, majRev, minRev)
as
(
    --Gets files with highest major version and all minor versions
    select fileId, majRev, minrev
    from files
    inner join latestFileMajorVersion on files.v_fileid = latestFileMajorVersion.fileId
    where v_majrev = latestFileMajorVersion.majRev
),
latestFileVersion(fileId, majRev, minRev)
as
(
    select fileId, majRev, minrev
    from latestFileVersions
    where --I'm stuck here.
)
select v_fileid, v_majrev, v_minrev, v_status
from files
inner join latestFileVersion on v_fileid = latestFileVersion.fileId
where latestFileVersion.majRev = v_majrev and latestFileVersion.minRev = v_minrev and v_status = 'UnAssigned'

表格：

CREATE TABLE [dbo].[Files](
    [v_fileid] [uniqueidentifier] NOT NULL,
    [v_libid] [uniqueidentifier] NOT NULL,
    [v_userid] [uniqueidentifier] NOT NULL,
    [v_majrev] [int] NOT NULL,
    [v_minrev] [int] NOT NULL,
    [v_status] [nvarchar](16) NOT NULL,
 CONSTRAINT [PK_Files] PRIMARY KEY CLUSTERED 
(
    [v_fileid] ASC,
    [v_majrev] ASC,
    [v_minrev] ASC
)WITH (PAD_INDEX  = OFF, STATISTICS_NORECOMPUTE  = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS  = ON, ALLOW_PAGE_LOCKS  = ON) ON [PRIMARY],
 CONSTRAINT [IX_Files] UNIQUE NONCLUSTERED 
(
    [v_majrev] ASC
)WITH (PAD_INDEX  = OFF, STATISTICS_NORECOMPUTE  = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS  = ON, ALLOW_PAGE_LOCKS  = ON) ON [PRIMARY],
 CONSTRAINT [IX_Files] UNIQUE NONCLUSTERED 
(
    [v_minrev] ASC
)WITH (PAD_INDEX  = OFF, STATISTICS_NORECOMPUTE  = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS  = ON, ALLOW_PAGE_LOCKS  = ON) ON [PRIMARY]
) ON [PRIMARY]

Answer 1

您可以使用此构思重写查询。它适用于您提供的样本数据。

declare @t table (fileid varchar(10), [version] varchar(5))

insert @t values('FileA','1.0')
insert @t values('FileA','1.1')
insert @t values('FileA','1.2')
insert @t values('FileA','1.3')
insert @t values('FileB','2.1')
insert @t values('FileB','2.2')
insert @t values('FileB','2.3')
insert @t values('FileB','2.4')
insert @t values('FileC','5.1')
insert @t values('FileC','5.2')
insert @t values('FileC','5.3')

SELECT * FROM @t t 
WHERE NOT EXISTS 
(SELECT 1 FROM @t WHERE PARSENAME([version], 1) = '0' AND t.fileid = fileid)

Answer 2

这是一个可以使用的小示例表的工作解决方案。完全按原样执行它，周围没有其他sql。然后，您可以根据需要对其进行修改。

declare @files table (fileId int identity(1,1), majRev int, minRev int)
insert into @files values (1,0)
insert into @files values (1,1)
insert into @files values (2,0)
insert into @files values (2,2)
insert into @files values (3,1)
insert into @files values (3,2)
insert into @files values (4,1)

 select fileId, majRev, minRev
   from @files f
  where not exists 
        (select *
           from @files
          where minRev = 0
            and majRev = f.majRev
            and fileId = f.fileId)

Answer 3

当我尝试编写某些内容时，有时候我会过于专注于制作一个“优雅”的解决方案，而不是仅仅强制解决方案并将其应用于特定问题。

所以我想知道是否1）我已经强行推行这个解决方案2）是否有更“优雅”或聪明的方式去写这个。欢迎所有评论，批评和/或建议。

;with solution(fileId, majrev, minrev, status, checkid, maxrev) 
as
(
select *
from files
inner join
(
    select v_fileid, MAX(v_majrev) as max_major
    from files
    group by v_fileid
) as max_versions on max_versions.v_fileid = files.v_fileid and 
                     max_major = files.v_majrev
except                   
select *
from files
inner join
(
    select v_fileid, MAX(v_majrev) as max_major
    from files
    where v_minrev = 0
    group by v_fileid
) as unwanted_versions on unwanted_versions.v_fileid = files.v_fileid and 
                          max_major = files.v_majrev
)
select *
from solution

使用T-SQL中的两列过滤数据

3 个答案: