I have this query:
CREATE TABLE [factOffertDetail](
[idOffertRow] [INT] IDENTITY(1,1) NOT NULL,
[idOffertRegion] [INT] NOT NULL,
[idProduct] [INT] NOT NULL,
[Qty] [DECIMAL](12, 2) NULL,
[idUnitPrice] [TINYINT] NULL
)
DECLARE #TMP2 TABLE (
idOffertRowNEW INT,
idOffertRow INT
)
INSERT INTO factOffertDetail
( idOffertRegion ,
idProduct ,
Qty ,
idUnitPrice
)
OUTPUT inserted.idOffertRow INTO #TMP2(d.idOffertRowNEW)
SELECT
d.idOffertRegion,
d.idProduct ,
d.Qty ,
d.idUnitPrice
FROM factOffertDetail d
I need to get the keys of the old and the new idOffertRow generated by identity.
idOffertRow is the identity (1,1) key of the factOffertDetail table.
How can I do this with an insert ?
Is it possible or I have to switch to merge command ?
Thanks to support
I would recommend to doing this:
Alter your table with new coloum,
ALTER TABLE [factOffertDetail]
ADD [ParentId] [INT] NULL
then,
INSERT INTO factOffertDetail
( ParentId,
idOffertRegion ,
idProduct ,
Qty ,
idUnitPrice
)
OUTPUT inserted.idOffertRow,inserted.ParentId INTO #TMP2(idOffertRowNEW,idOffertRow)
SELECT
d.idOffertRow,
d.idOffertRegion,
d.idProduct ,
d.Qty ,
d.idUnitPrice
FROM factOffertDetail d
Thank You!
Related
I have a user defined type which is passed to my sql stored procedure:
CREATE TYPE [dbo].[LearningDelivery] AS TABLE(
[LearnAimRef] NVARCHAR(10) NOT NULL,
[AimType] INT NOT NULL,
[AimSeqNumber] INT NOT NULL,
);
i have a table with structure as below. I want to match join the type name from the above definition into the ElementName column of the FieldDefinitation table and insert separate insert statements. can anyone help me how to achieve this?
CREATE TABLE [dbo].[LearningAimData]
(
[Id] INT IDENTITY(1, 1) NOT NULL,
[LearnerId] INT NOT NULL,
[FieldId] INT NOT NULL,
[Data] VARCHAR(128) NOT NULL
)
I want to insert each type i.e LearnAimRef , AimType as individual row values in the LearningAimData.
I have a lookup table for [FieldId]
CREATE TABLE [dbo].[FieldDefinition]
(
[Id] INT IDENTITY(1, 1),
[FieldId] INT NOT NULL,
[Name] VARCHAR(50) NOT NULL,
[ElementName] VARCHAR(50) NOT NULL
)
I use bellow code to archive old data in ArchiveTable and delete archived data from SourceTable
DELETE FROM SourceTable
OUTPUT
DELETED.[ID],
DELETED.[Code],
DELETED.[Title]
INTO ArchiveTable([OldID], [Code], [Title])
WHERE Condition
Structure of tables:
CREATE TABLE [SourceTable](
[ID] [INT] IDENTITY(1,1) NOT NULL,
[Code] [VARCHAR](16) NULL,
[Title] [NVARCHAR](128) NULL,
CONSTRAINT [PK_SourceTable] PRIMARY KEY CLUSTERED ([ID] ASC)
)
GO
CREATE TABLE [ArchiveTable](
[ID] [INT] IDENTITY(1,1) NOT NULL,
[OldID] [INT] NOT NULL,
[Code] [VARCHAR](16) NULL,
[Title] [NVARCHAR](128) NULL,
CONSTRAINT [PK_ArchiveTable] PRIMARY KEY CLUSTERED ([ID] ASC)
)
GO
I need to return deleted records and ArchiveTable.[ID] to application. I change the code like this:
DELETE FROM SourceTable
OUTPUT
DELETED.[ID],
DELETED.[Code],
DELETED.[Title]
INTO ArchiveTable([OldID], [Code], [Title])
OUTPUT DELETED.*
WHERE Condition
This code return deleted records but I don't know how to get ID of ArchiveTable for this records. Look at ArchiveTable structure, It has OldID column that refer to SourceTable.ID and ID column that it is an Identity column of ArchiveTable. I need to ArchiveTable.ID in final result.
You can use a temporary table
CREATE TABLE #DeletedRows(
[ID] [INT] NOT NULL,
[Code] [VARCHAR](16) NULL,
[Title] [NVARCHAR](128) NULL
)
DELETE SourceTable
OUTPUT
DELETED.[ID],
DELETED.[Code],
DELETED.[Title]
INTO #DeletedRows([ID], [Code], [Title])
WHERE Condition
INSERT ArchiveTable([OldID], [Code], [Title])
OUTPUT INSERTED.*
SELECT [ID], [Code], [Title]
FROM #DeletedRows
DROP TABLE #DeletedRows
A variant with a table variable
DECLARE #DeletedRows TABLE(
[ID] [INT] NOT NULL,
[Code] [VARCHAR](16) NULL,
[Title] [NVARCHAR](128) NULL
)
DELETE SourceTable
OUTPUT
DELETED.[ID],
DELETED.[Code],
DELETED.[Title]
INTO #DeletedRows([ID], [Code], [Title])
WHERE Condition
INSERT ArchiveTable([OldID], [Code], [Title])
OUTPUT INSERTED.*
SELECT [ID], [Code], [Title]
FROM #DeletedRows
I found an interesting variant using DML with OUTPUT in SP and INSERT...EXEC... after that:
Test tables:
CREATE TABLE TestTable(
ID int NOT NULL PRIMARY KEY,
Title varchar(10) NOT NULL
)
CREATE TABLE TestTableLog(
LogID int NOT NULL IDENTITY,
OperType char(1) NOT NULL,
CHECK(OperType IN('I','U','D')),
ID int NOT NULL,
Title varchar(10) NOT NULL
)
DML procedures:
CREATE PROC InsTestTable
#ID int,
#Title varchar(10)
AS
INSERT TestTable(ID,Title)
OUTPUT inserted.ID,inserted.Title,'I' OperType
VALUES(#ID,#Title)
GO
CREATE PROC UpdTestTable
#ID int,
#Title varchar(10)
AS
UPDATE TestTable
SET
Title=#Title
OUTPUT inserted.ID,inserted.Title,'U' OperType
WHERE ID=#ID
GO
CREATE PROC DelTestTable
#ID int
AS
DELETE TestTable
OUTPUT deleted.ID,deleted.Title,'D' OperType
WHERE ID=#ID
GO
Tests:
-- insert test
INSERT TestTableLog(ID,Title,OperType)
EXEC InsTestTable 1,'A'
INSERT TestTableLog(ID,Title,OperType)
EXEC InsTestTable 2,'B'
INSERT TestTableLog(ID,Title,OperType)
EXEC InsTestTable 3,'C'
-- update test
INSERT TestTableLog(ID,Title,OperType)
EXEC UpdTestTable 2,'BBB'
-- delete test
INSERT TestTableLog(ID,Title,OperType)
EXEC DelTestTable 3
GO
-- show resutls
SELECT *
FROM TestTableLog
Maybe it'll be interesting to someone.
I have a table called tblPacks.
CREATE TABLE [dbo].[tblPacks]
(
[ID] [int] NOT NULL,
[BatchNumber] [varchar](30) NULL,
[PackID] VARCHAR(50),
[Status] [int] NULL
)
And a stored procedure spInsertPacks.
CREATE PROCEDURE spInsertPacks
#ID INT,
#BatchNumber VARCHAR(30),
#Count INT
AS
BEGIN
INSERT INTO tblPacks
Values
(
#ID,
#BatchNumber,
CONVERT([varchar](50),
'PK'+
case
when len(#ID)<=(3) then CONVERT([varchar](20),right((0.001)*#ID,(3)),0)
else CONVERT([varchar](20),#ID,0)
end,0),0)
END
If ID of data type INT inserted in an order like 1,2,3,4,5... the above logic works fine. But there is no restriction for a user to enter random numbers. I want a stored procedure to generate PackID(PK001,PK002..) sequence in order, irrespective of #ID and ID. Cannot be an identity Column. How can I do that?
Actually This PackID is a barcode If barcode already existed for Pack then that sequence may not be same with the sequence we used and Newly generated barcodes which we are generating will be in seuquence PK001
Sample Output:-
ID BatchNumber PackID Status
1 b1 PK001 0
1 b2 Pk002 0
5 b7 ABC768 0
3 b2 PK003 0
I have simplified the logic a bit for generating PackID
Add a new column(identifier) for identifying the code and use it for PackID generation and for sequence use Identity column
CREATE TABLE [dbo].[tblPacks]
(
Iden_ID INT IDENTITY(1, 1),
[ID] [INT] NOT NULL,
[BatchNumber] [VARCHAR](30) NULL,
[Identifier] [VARCHAR](50),
[PackID] AS [Identifier]
+ CASE
WHEN Iden_ID <= 999 THEN RIGHT('00' + CONVERT(VARCHAR(3), ID), 3)
ELSE CONVERT([VARCHAR](20), ID, 0)
END,
[Status] [INT] NULL
)
To check the working
INSERT INTO [dbo].[tblPacks]
([ID],identifier,[BatchNumber],[Status])
VALUES (1,'pk','bat',1)
SELECT *
FROM [tblPacks]
I am using Microsoft SQL Server Management Studio. So this is what I have right now:
CREATE TABLE Product
(
product_id INT NOT NULL PRIMARY KEY IDENTITY,
product_code CHAR(4) NOT NULL, --For a book use 'BOOK'
product_name VARCHAR(40) NOT NULL,
product_desc VARCHAR(5000),
book_author INT,
book_publisher INT,
product_price SMALLMONEY NOT NULL CHECK (product_price >= 0),
FOREIGN KEY (book_author) REFERENCES Author
);
So I would like to make it so that book_author and book_publisher cannot be null if product_code == 'BOOK'.
Is this possible and how?
This is how I would do it:
CREATE TABLE Product (
product_id INT NOT NULL PRIMARY KEY IDENTITY
, product_code CHAR(4) NOT NULL
, product_name VARCHAR(40) NOT NULL
, product_desc VARCHAR(5000)
, book_author INT
, book_publisher INT
, product_price SMALLMONEY NOT NULL CHECK (product_price >= 0)
, CONSTRAINT CHK_author CHECK (
CASE
WHEN product_code = 'BOOK' AND (book_author IS NULL OR book_publisher IS NULL) THEN 0
ELSE 1
END = 1
)
);
CHK_author constraint will check if your product_code is BOOK, and if it is, then it will check whether book_author OR book_publisher are NULL values. If one of them is - it will restrict statement.
Here's a SQL Fiddle
CREATE TABLE [dbo].[Product](
[product_id] [int] IDENTITY(1,1) NOT NULL,
[product_code] [char](4) NOT NULL,
[product_name] [varchar](40) NOT NULL,
[product_desc] [varchar](5000) NULL,
[book_author] [int] NULL,
[book_publisher] [int] NULL,
[product_price] [smallmoney] NOT NULL,
PRIMARY KEY CLUSTERED ( [product_id] ASC )
)
Go
ALTER TABLE [dbo].[Product] WITH CHECK ADD CONSTRAINT [CK_Product] CHECK (
(
( [product_code] <> 'BOOK' )
OR
(
( [book_author] IS NOT NULL )
AND ( [book_publisher] IS NOT NULL )
)
)
AND
( [product_price] >= 0 )
)
Go
ALTER TABLE [dbo].[Product] CHECK CONSTRAINT [CK_Product]
Go
One constraint pair row.
I have the following three tables (each RobotPart has exactly one arm and one leg). I have a large number of {arm, leg} pairs need to be inserted. Any new combinations of arm and leg will create a new entry in RobotPart. Any existing combination will not be inserted. No updates are needed for either parent or children. I need preserve the identity columns. Any efficient ways to accomplish this in SQL?
CREATE TABLE [dbo].[Arm](
[Id] [int] IDENTITY(1,1) NOT NULL,
[Model] [varchar](20) NULL,
CONSTRAINT [PK_dbo.Arm] PRIMARY KEY CLUSTERED
(
[Id] ASC
)
) ON [PRIMARY]
CREATE TABLE [dbo].[Leg](
[Id] [int] IDENTITY(1,1) NOT NULL,
[Model] [varchar](10) NULL,
CONSTRAINT [PK_dbo.Leg] PRIMARY KEY CLUSTERED
(
[Id] ASC
)
) ON [PRIMARY]
CREATE TABLE [dbo].[RobotPart](
[Id] [int] IDENTITY(1,1) NOT NULL,
[Name] [varchar](20) NOT NULL,
[ArmId] [int] NOT NULL,
[LegId] [int] NOT NULL,
CONSTRAINT [PK_dbo.RobotPart] PRIMARY KEY CLUSTERED
(
[Id] ASC
)
) ON [PRIMARY]
GO
ALTER TABLE [dbo].[RobotPart] WITH CHECK ADD CONSTRAINT [FK_dbo.RobotPart_dbo.Arm_ArmId] FOREIGN KEY([ArmId])
REFERENCES [dbo].[Arm] ([Id])
GO
ALTER TABLE [dbo].[RobotPart] WITH CHECK ADD CONSTRAINT [FK_dbo.RobotPart_dbo.Leg_LegId] FOREIGN KEY([LegId])
REFERENCES [dbo].[Leg] ([Id])
GO
Step 1: Insert new arms and legs using a LEFT OUTER JOINs from your source table to the arms and legs table, respectively. (a separate insert statement for each table)
Step 2: Insert the new combinations using an inner join from your source table to arms and legs and a left outer join from the your source table to the RobotPart table.
Here is my initial try on it using MERGE. Not sure how this compare to Lmu92 proposed.
CREATE TYPE [dbo].[RobotPart_udtt] AS TABLE(
[Arm] [varchar](20) NOT NULL,
[Leg] [varchar](10) NOT NULL,
[Name] [varchar](20) NOT NULL
)
GO
CREATE PROCEDURE dbo.[prc_Component_Create]
#robotParts [RobotPart_udtt] READONLY
AS
BEGIN
SET NOCOUNT ON;
DECLARE #messageId INT
DECLARE #status INT
MERGE [Arm] AS TARGET
USING (
SELECT
tR.arm AS Model
FROM #robotParts AS tR
) AS SOURCE
ON TARGET.Model = SOURCE.Model
WHEN NOT MATCHED THEN
INSERT
(
Model
)
VALUES
(
SOURCE.Model
);
MERGE [Leg] AS TARGET
USING (
SELECT
tR.leg AS Model
FROM #robotParts AS tR
) AS SOURCE
ON TARGET.Model = SOURCE.Model
WHEN NOT MATCHED THEN
INSERT
(
Model
)
VALUES
(
SOURCE.Model
);
WITH NewParts (ArmId, LegId, Name)
AS
(
SELECT tA.Id
, tL.Id
, tR.Name
FROM #robotParts AS tR
INNER JOIN [Arm] AS tA
ON tR.Arm = tA.Model
INNER JOIN [Leg] AS tL
ON tR.Leg = tL.Model
)
INSERT INTO RobotPart (Name, ArmId, LegId)
SELECT tN.Name
, tN.ArmId
, tN.LegId
FROM NewParts AS tN
LEFT JOIN RobotPart AS tR
ON tR.ArmId = tN.ArmId
AND tR.LegId = tN.LegId
WHERE tR.ArmId IS NULL AND tR.LegId IS NULL
END