I am not sure if my title really explains the question, so I'll try an example:
Let's say I have:
1) region table: parent regions and subregions in the same table (link A)
2) product table: each product is linked to a single parent region (link B)
3) product_price table: lists the price of a product (C) in all sub regions of that product region (link D >>> the link in question).
(the diagram is showing only the relevant fields, there are a lot more data in region and product)
Is there a way to define the region->region_price key, to include only sub_region_id's of the parent_region in the product->region key??
Or, in the terms of the image, how do I make the D link to include only regions that are children [as in A] of the B & C link?
Hope you are getting my point...
Here are the real tables and links:
CREATE TABLE [dbo].[product](
[product_id] [int] NOT NULL,
[product_name] [nchar](10) NOT NULL,
[parent_region_id] [int] NOT NULL,
CONSTRAINT [PK_product] PRIMARY KEY CLUSTERED
(
[product_id] 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].[product_price](
[product_id] [int] NOT NULL,
[sub_region_id] [int] NOT NULL,
[price] [decimal](18, 0) NOT NULL,
CONSTRAINT [PK_product_price] PRIMARY KEY CLUSTERED
(
[product_id] ASC,
[sub_region_id] 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].[region](
[region_id] [int] NOT NULL,
[region_name] [nvarchar](50) NOT NULL,
[parent_region_id] [int] NULL,
CONSTRAINT [PK_region] PRIMARY KEY CLUSTERED
(
[region_id] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]
ALTER TABLE [dbo].[product] WITH CHECK ADD CONSTRAINT [FK_product_region] FOREIGN KEY([parent_region_id])
REFERENCES [dbo].[region] ([region_id])
ALTER TABLE [dbo].[product] CHECK CONSTRAINT [FK_product_region]
ALTER TABLE [dbo].[product_price] WITH CHECK ADD CONSTRAINT [FK_product_price_product] FOREIGN KEY([product_id])
REFERENCES [dbo].[product] ([product_id])
ALTER TABLE [dbo].[product_price] CHECK CONSTRAINT [FK_product_price_product]
ALTER TABLE [dbo].[product_price] WITH CHECK ADD CONSTRAINT [FK_product_price_region] FOREIGN KEY([sub_region_id])
REFERENCES [dbo].[region] ([region_id])
ALTER TABLE [dbo].[product_price] CHECK CONSTRAINT [FK_product_price_region]
ALTER TABLE [dbo].[region] WITH CHECK ADD CONSTRAINT [HK_region_region] FOREIGN KEY([parent_region_id])
REFERENCES [dbo].[region] ([region_id])
ALTER TABLE [dbo].[region] CHECK CONSTRAINT [HK_region_region]
I guess you are trying to enforce a business rule with a database rule, but they're not always the same. Instead, you can run validation querys in your code before inserting or updating, or you could implement a trigger that validates your rules or throws an error.
Related
Would someone help, please, to get rid of the error:
Violation of PRIMARY KEY constraint 'PK_stmp_tst1'. Cannot insert duplicate key in object 'dbo.stmp_tst'. The duplicate key value is (1).
which occurs after switch partition of the table.
Full SQL Script below:
I. Create 2 equal tables in different schemas:
CREATE TABLE dbo.stmp_tst(
[inn] [varchar](20) NULL,
[id] [bigint] IDENTITY(1,1) NOT NULL,
CONSTRAINT [PK_stmp_tst] PRIMARY KEY CLUSTERED
(
[id] 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 stage.stmp_tst(
[inn] [varchar](20) NULL,
[id] [bigint] IDENTITY(1,1) NOT NULL,
CONSTRAINT [PK_stmp_tst] PRIMARY KEY CLUSTERED
(
[id] 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
II. Insert data into stage table.
insert into stage.stmp_tst (inn)
select '1111'
III. Switch-partition stage to dbo.
alter table stage.stmp_tst switch partition 1 to dbo.stmp_tst partition 1;
IV. Add data to dbo table.
insert into dbo.stmp_tst (inn)
select '1111'
V. We have the error:
Violation of PRIMARY KEY constraint 'PK_stmp_tst1'. Cannot insert
duplicate key in object 'dbo.stmp_tst'. The duplicate key value is
(1).
IV. Drop temporary tables:
drop table dbo.stmp_tst
drop table stage.stmp_tst
It can be solved by the query:
DBCC CHECKIDENT ('dbo.stmp_tst', RESEED);
but reseeding takes time.
Is it possible to do a switch-partition correctly without reseed?
Thank you.
I have three tables: offices, suboffices, and sales. Each office owns several suboffices, both offices and suboffices sell products. How should I design my sales table to store there office or suboffice where this sale was made?
I was thinking about having a compound foreign key made of office_id and suboffice_id (where suboffice_id may be null, in which case a sale was made in office)
Is it the right way to design a database?
I was also thinking about having two sales tables: for offices and suboffices. But in my opinion it makes things a bit harder..
UPDATE
Sales can be of different types, so they will need different tables. And there will be one table 'SALES' which will store type of a sale and where it was made
Your sub-offices sound the same as your offices from a properties point of view. Here is a diagram of what your table design could look like by having your Office table include a ParentOffice foreign key.
And here is the SQL to create those tables:
CREATE TABLE [dbo].[Office](
[OfficeId] [int] NOT NULL,
[ParentOfficeId] [int] NULL,
[MoreStuff] [nvarchar](50) NULL,
CONSTRAINT [PK_Office] PRIMARY KEY CLUSTERED
(
[OfficeId] 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
ALTER TABLE [dbo].[Office] WITH CHECK ADD CONSTRAINT [FK_Office_Office] FOREIGN KEY([ParentOfficeId])
REFERENCES [dbo].[Office] ([OfficeId])
GO
ALTER TABLE [dbo].[Office] CHECK CONSTRAINT [FK_Office_Office]
GO
CREATE TABLE [dbo].[Product](
[ProductId] [int] NOT NULL,
[MoreStuff] [nvarchar](50) NULL,
CONSTRAINT [PK_Product] PRIMARY KEY CLUSTERED
(
[ProductId] 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].[Sale](
[SaleId] [int] NOT NULL,
[ProducitId] [int] NOT NULL,
[OfficeId] [int] NOT NULL,
CONSTRAINT [PK_Sale] PRIMARY KEY CLUSTERED
(
[SaleId] 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
ALTER TABLE [dbo].[Sale] WITH CHECK ADD CONSTRAINT [FK_Sale_Office] FOREIGN KEY([OfficeId])
REFERENCES [dbo].[Office] ([OfficeId])
GO
ALTER TABLE [dbo].[Sale] CHECK CONSTRAINT [FK_Sale_Office]
GO
ALTER TABLE [dbo].[Sale] WITH CHECK ADD CONSTRAINT [FK_Sale_Sale] FOREIGN KEY([ProducitId])
REFERENCES [dbo].[Product] ([ProductId])
GO
ALTER TABLE [dbo].[Sale] CHECK CONSTRAINT [FK_Sale_Sale]
GO
One way to go about this, assuming OFFICE and SUBOFFICE share many of the same attributes is to eliminate the SUBOFFICE table and add a PARENT_OFFICE_ID column to the OFFICE table. Under this design, the distinction between an office and a sub-office would be whether or not PARENT_OFFICE_ID is null. Then SALES can simply have an OFFICE_ID column that can reference either kind of office.
I have two tables. First table has multi columns as the primary key and the second table has one column primary key. Data has be entered first in Table1 FieldPlacement where FieldPlacementNum will be generated then enter a record in Table2 where the FieldPlacementNum has to exist in Table1.
Currently has one to many relationship but I want the reverse the relationship of tables and SQL does not let me do it. thanks
Table 1
CREATE TABLE [dbo].[FieldPlacement]
(
[ID] [varchar](10) NOT NULL,
[Year] [varchar](10) NOT NULL,
[Term] [varchar](10) NOT NULL,
[PlacementNum] [int] IDENTITY(1,1) NOT NULL,
[Email] [varchar](70) NULL,
CONSTRAINT [PK_FieldPlacement]
PRIMARY KEY CLUSTERED ([ID] ASC, [Year] ASC, [Term] ASC, [PlacementNum] 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
SET ANSI_PADDING OFF
GO
ALTER TABLE [dbo].[FieldPlacement] WITH CHECK
ADD CONSTRAINT [FK_FieldPlacement_FieldPlacementEval1]
FOREIGN KEY([PlacementNum])
REFERENCES [dbo].[FieldPlacementEval] ([PlacementNum])
GO
ALTER TABLE [dbo].[FieldPlacement]
CHECK CONSTRAINT [FK_FieldPlacement_FieldPlacementEval1]
GO
Table 2
CREATE TABLE [dbo].[FieldPlacementEval]
(
[PlacementNum] [int] NOT NULL,
[StudentLastName] [varchar](50) NULL,
[StudentFirstName] [varchar](50) NULL,
[TeacherLastName] [varchar](50) NULL,
[TeacherFirstName] [varchar](50) NULL,
CONSTRAINT [PK_FieldPlacementEval]
PRIMARY KEY CLUSTERED ([PlacementNum] ASC)
WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF,
IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON,
ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]
Since FieldPlacement.PlacementNum is an IDENTITY, which is NOT NULL and unique, you could also make this your PK, and then just create a unique index on those four columns that currently make up the PK (to ensure their uniqueness).
From your FieldPlacementEval just reference the PlacementNum.
The additional benefit here would be a smaller, more efficient clustering key on your FieldPlacement table (since it's only 1 column instead of 4 - and skips all those variable length columns which are really bad for a clustering key)
I have three tables a,b,and c each with an int IDENTITY PK field that relates to the child table.
CREATE TABLE [dbo].[a]([aID] [int] IDENTITY(1,1) NOT NULL,[aCode] [varchar](20) NOT NULL,
CONSTRAINT [PK_a] PRIMARY KEY CLUSTERED([aID] 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].[b]([bID] [int] IDENTITY(1,1) NOT NULL,[aID] [int] NOT NULL, [bCode] [varchar](20) NOT NULL,
CONSTRAINT [PK_b] PRIMARY KEY CLUSTERED([bID] 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].[c]([cID] [int] IDENTITY(1,1) NOT NULL,[bID] [int] NOT NULL,[cCode] [varchar](20) NOT NULL,
CONSTRAINT [PK_c] PRIMARY KEY CLUSTERED([cID] 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
ALTER TABLE [dbo].[b] WITH CHECK ADD CONSTRAINT [FK_b_a] FOREIGN KEY([aID])
REFERENCES [dbo].[a] ([aID])
GO
ALTER TABLE [dbo].[b] CHECK CONSTRAINT [FK_b_a]
GO
ALTER TABLE [dbo].[c] WITH CHECK ADD CONSTRAINT [FK_c_b] FOREIGN KEY([bID])
REFERENCES [dbo].[b] ([bID])
GO
ALTER TABLE [dbo].[c] CHECK CONSTRAINT [FK_c_b]
GO
How do I create a CONSTRAINT that enforces a unique condition for a.aID, c.cCode?
If you are looking for a unique constraint over a combination of columns a.aID, c.cCode one way is to alter table c and add [aid] column to it and have a composite unique key.
ALTER TABLE [dbo].[c](
[cID] [int] IDENTITY(1,1) NOT NULL,
[bID] [int] NOT NULL,[cCode] [varchar](20) NOT NULL,
[cCode] [varchar](20) NOT NULL,
[aID] INT FOREIGN KEY REFERENCES [a]([aID]) NOT NULL,
CONSTRAINT uq_cCode_aid UNIQUE NONCLUSTERED (cCode,aID),
CONSTRAINT [PK_c] PRIMARY KEY CLUSTERED([cID] 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
You can add a unique constraint as follows:
ALTER TABLE [C] ADD CONSTRAINT [uc_C_cCode] UNIQUE NONCLUSTERED [cCode];
You cannot however enforce a constraint across tables, which is what it sounds like you are trying to do.
As for table A.aID, its a primary key, it will already have a unique constraint.
I found a way to get what I wanted (I think) using an indexed view. Please comment on if this seems appropriate. My tests so far conclude that it does work.
I created a view with schemabinding on joining c to a (going through b).
I created an clustered index to enforce unique on aID and cCode.
This allows me to not need to include aID within the c table but still enforces the unique constraint.
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
CREATE VIEW [dbo].[vw.aIDcCode] WITH SCHEMABINDING
AS
SELECT dbo.a.aID, dbo.c.cCode
FROM dbo.a
INNER JOIN
dbo.b ON dbo.a.aID = dbo.b.aID
INNER JOIN
dbo.c ON dbo.b.bID = dbo.c.bID
GO
SET ARITHABORT ON
GO
SET CONCAT_NULL_YIELDS_NULL ON
GO
SET QUOTED_IDENTIFIER ON
GO
SET ANSI_NULLS ON
GO
SET ANSI_PADDING ON
GO
SET ANSI_WARNINGS ON
GO
SET NUMERIC_ROUNDABORT OFF
GO
CREATE UNIQUE CLUSTERED INDEX [IDX_Unique_aIDcCode] ON [dbo].[vw.aIDcCode]
([aID] ASC, [cCode] ASC)
WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, SORT_IN_TEMPDB = OFF, IGNORE_DUP_KEY = OFF, DROP_EXISTING = OFF, ONLINE = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON)
ON [PRIMARY]
GO
I haven't ever really used foreign keys in my databases, I have always just written code that would enforce things for me, but I'm looking towards shifting some of that logic to the database and learning more about foreign keys.
I currently have a database which has the table Categories:
CREATE TABLE [dbo].[Categories](
[CategoryID] [bigint] IDENTITY(1,1) NOT NULL,
[ParentCategoryID] [bigint] NOT NULL,
[CategoryStatus] [bit] NOT NULL,
[CategoryTitle] [varchar](64) NOT NULL,
[CategorySlug] [varchar](64) NOT NULL,
[CategoryThumbnail] [varchar](128) NULL,
[CategoryHeaderImage] [varchar](128) NULL,
[CategoryDescription] [text] NULL,
CONSTRAINT [PK_Categories] PRIMARY KEY CLUSTERED
(
[CategoryID] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY],
CONSTRAINT [UI_Categories_Slug] UNIQUE NONCLUSTERED
(
[CategorySlug] 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]
I would like to create a foreign key that relates ParentCategoryID back up to CategoryID. When creating the FK, which would be the primary column, and which would be the foreign column in this instance?
I would think that ParentCategoryID would have to be NULLable (unless you want a top-level parent to point to itself, and that doesn't make much sense).
ALTER TABLE dbo.Categories
ADD CONSTRAINT FK_SelfParent
FOREIGN KEY (ParentCategoryID)
REFERENCES dbo.Categories(CategoryID);
True story: at a job interview at Microsoft (several years ago), a SQL Server person told me that this wasn't possible.