问题问题与SSIS包
在经过几个月下面描述的过程没有任何问题已经工作加载数据时,了解间歇不一致绝大多数已运行时间(2008 R2)。然而,我们有三个错误连接的数据。问题是,是什么造成了这种情况,我该如何补救?
DATA_PreImp
sourceid col01 col02 col03 col04 col...
100001 John Smith
100002 Calvin Klein
100003 Peter Parker
100004 Moe Greene
一般所呈现的结果是该属性被连接到Items_Main正确,但有时(小于1%)的顺序进行加扰,使得COL01的值不被连接到相同的Items_Main为的值其余的列。
任何有关这是什么导致的见解将不胜感激。
的数据移动过程
我们有一个SSIS包从称为DATA_PreImp到(基于属性)由三个相关的表的结构的平坦表传输数据。
- Items_Main应该包含一个行用于DATA_PreImp每一行
- Items_Featurevalues包含用于DATA_PreImp一个行的每个列值一行
- Items_MainRel包含Items_Main和Items_FeatureValues
的之间的连接SSIS包中的第一步是将DATA_PreImp中的数据插入到Items_Main中,并将生成的标识符插入到空DATA_PreImpMappingTMP表中的TARGET_ID列中。
insert into items_main(creationdate, status)
output inserted.itemid into DATA_PreImpMappingTMP(TARGET_ID)
select getdate(), '0' from data_preimp
order by sourceid asc;
SSIS包第二步填充TARGET_ID(ITEMID最初)的Items_MainRel表和用于特征的标识符(在这种情况下为5)。
insert into items_mainrel(itemid, featureid)
output inserted.itemrelid into DATA_PreImpMapping2TMP(INDREL_ID)
select TARGET_ID, 5 from DATA_PreImpMappingTMP
order by TARGET_ID asc;
的第三步是从与所述DATA_PreImp SOURCE_ID填写DATA_PreImpMapping2TMP表中的列SOURCE_ID。
with cte as (select sourceid, row_number() over (order by sourceid asc) as row from data_preimp)
update m set m.SOURCE_ID = s.sourceid, m.FEAT_ID = 5
from DATA_PreImpMapping2TMP as m
join cte as s on s.row = m.ROW;
的最后一步是填充Items_FeatureValues表与来自DATA_PreImpMapping2TMP和DATA_PreImp数据。
insert into items_featurevalues(itemrelid, languageid, fnvarchar)
select DATA_PreImpMapping2TMP.INDREL_ID, 0, data_preimp.col01
from DATA_PreImpMapping2TMP
join data_preimp on (DATA_PreImpMapping2TMP.SOURCE_ID = data_preimp.sourceid)
where FEAT_ID = 5
数据表结构
这里是什么是需要创建方案:
CREATE TABLE [dbo].[DATA_PreImp](
[sourceid] [bigint] IDENTITY(1,1) NOT NULL,
[col01] [nvarchar](500) NULL,
[col02] [nvarchar](500) NULL,
[col03] [nvarchar](500) NULL,
[col04] [nvarchar](500) NULL,
[col05] [nvarchar](500) NULL,
[col06] [nvarchar](500) NULL,
[col07] [nvarchar](500) NULL,
[col08] [nvarchar](500) NULL,
[col09] [nvarchar](500) NULL,
[col10] [nvarchar](500) NULL,
CONSTRAINT [PK_DATA_PreImp] PRIMARY KEY CLUSTERED
(
[sourceid] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]
CREATE TABLE [dbo].[DATA_PreImpMappingTMP](
[ROW] [int] IDENTITY(1,1) NOT NULL,
[TARGET_ID] [int] NULL,
PRIMARY KEY CLUSTERED
(
[ROW] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]
CREATE TABLE [dbo].[Items_Main](
[Itemid] [int] IDENTITY(1,1) NOT NULL,
[creationDate] [smalldatetime] NOT NULL,
[status] [int] NOT NULL,
[purchdate] [smalldatetime] NULL,
[logindate] [smalldatetime] NULL,
CONSTRAINT [PK_Items_Main] PRIMARY KEY CLUSTERED
(
[Itemid] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]
CREATE TABLE [dbo].[DATA_PreImpMapping2TMP](
[ROW] [int] IDENTITY(1,1) NOT NULL,
[SOURCE_ID] [int] NULL,
[INDREL_ID] [int] NULL,
[FEAT_ID] [int] NULL,
PRIMARY KEY CLUSTERED
(
[ROW] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]
CREATE TABLE [dbo].[Items_Features](
[featureId] [int] IDENTITY(1,1) NOT NULL,
[featureRef] [varchar](15) NOT NULL,
[featureName] [varchar](50) NOT NULL,
[creationDate] [smalldatetime] NOT NULL,
[status] [int] NOT NULL,
[fieldType] [varchar](50) NOT NULL,
[featureType] [int] NOT NULL,
[featureDesc] [varchar](500) NULL,
CONSTRAINT [PK_Items_Features] PRIMARY KEY CLUSTERED
(
[featureId] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]) ON [PRIMARY]
CREATE TABLE [dbo].[Items_MainRel](
[ItemRelId] [int] IDENTITY(1,1) NOT NULL,
[Itemid] [int] NOT NULL,
[featureId] [int] NOT NULL,
CONSTRAINT [PK_Items_MainRel] PRIMARY KEY CLUSTERED
(
[ItemRelId] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]
GO
CREATE TABLE [dbo].[Items_FeatureValues](
[valueId] [int] IDENTITY(1,1) NOT NULL,
[ItemRelId] [int] NOT NULL,
[languageId] [int] NOT NULL,
[FnVarChar] [nvarchar](250) NULL,
[FInt] [int] NULL,
[FImage] [int] NULL,
[FNText] [ntext] NULL,
[FSmallDateTime] [smalldatetime] NULL,
CONSTRAINT [PK_Items_FeatureValues] PRIMARY KEY CLUSTERED
(
[valueId] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY] TEXTIMAGE_ON [PRIMARY]
GO
ALTER TABLE [dbo].[Items_MainRel] WITH CHECK ADD CONSTRAINT [FK_Items_MainRel_Items_Features] FOREIGN KEY([featureId])
REFERENCES [dbo].[Items_Features] ([featureId])
GO
ALTER TABLE [dbo].[Items_MainRel] CHECK CONSTRAINT [FK_Items_MainRel_Items_Features]
GO
ALTER TABLE [dbo].[Items_MainRel] WITH CHECK ADD CONSTRAINT [FK_Items_MainRel_Items_Main] FOREIGN KEY([Itemid])
REFERENCES [dbo].[Items_Main] ([Itemid])
GO
ALTER TABLE [dbo].[Items_MainRel] CHECK CONSTRAINT [FK_Items_MainRel_Items_Main]
GO
ALTER TABLE [dbo].[Items_FeatureValues] WITH CHECK ADD CONSTRAINT [FK_Items_FeatureValues_Items_MainRel] FOREIGN KEY([ItemRelId])
REFERENCES [dbo].[Items_MainRel] ([ItemRelId])
ON DELETE CASCADE
GO
ALTER TABLE [dbo].[Items_FeatureValues] CHECK CONSTRAINT [FK_Items_FeatureValues_Items_MainRel]
GO
INSERT INTO DATA_PreImp (col01,col02,col03,col04)
VALUES('John', 'Smith', '1964', 'NewYork'),
('Calvin', 'Klein', '1960', 'Washington D. C.'),
('Peter', 'Parker', '1974', 'Losangles'),
('Moe', 'Greene', '1928', 'Lasvegas')
INSERT INTO Items_Features (featureRef, featureName, creationDate, [status], fieldType, featureType, featureDesc)
VALUES ('firstname','First_Name', GETDATE(), 0, 'FnVarChar', 3, 'FirstName'),
('lastname','Last_Name', GETDATE(), 0, 'FnVarChar', 3, 'LastName'),
('Birth','Birth', GETDATE(), 0, 'FnVarChar', 3, 'Birth'),
('City','City', GETDATE(), 0, 'FnVarChar', 3, 'City')
我在第二个脚本上获得外键违规。 dbo中没有记录。具有featureID = 5的Items_Features(如硬编码到该脚本中)。应该是脚本中使用的另一个featureID还是dbo.Items_Features中缺少记录? – Aphillippe 2013-03-27 09:03:58
你在插入中使用交易吗? SSIS包中的错误处理是否会抛出任何有用的数据?是否有其他进程在同一个db/tables上同时运行,可能会导致并发问题? – Valkyrie 2013-04-24 14:01:59
这个复杂的东西,没有真正的包装和所有的内部工作,找到问题将是相当具有挑战性的。我们需要更多的信息...... – mrkb80 2013-05-07 17:33:25