Choosing a clustered index

Mon, 01 Jan 0001 00:00:00 +0000

There are two broad kinds of index in SQL Server: clustered and nonclustered.

Their structure is the same, a B-tree search tree.

A nonclustered index copies the index key data into a physical structure separate from the table. It is the equivalent of the index at the end of a book: it does not mix with the body of the book itself.
A clustered index structures the table. Physically, the table rows are stored at the leaf level of the tree. It is the equivalent of a dictionary, where the whole book is ordered on a sorted search key.

When to choose a clustered index

A clustered index shapes the table itself. A clustered table has the following properties:

Managing several auto-increments

Mon, 01 Jan 0001 00:00:00 +0000

How do you manage several auto-increments on a single table, when one column has to be numbered within the scope of another?

The classic example is a table holding a document number — an invoice number, say — numbered per branch office:

CREATE TABLE dbo.Invoice (
 BranchNumber TINYINT NOT NULL,
 InvoiceNumber INT NOT NULL,
 CONSTRAINT pk_invoice PRIMARY KEY (InvoiceNumber, BranchNumber)
);

Here is a procedure that does it:

CREATE PROCEDURE [dbo].[AddInvoice]
 @BranchNumber tinyint
AS BEGIN
 SET NOCOUNT ON

 DECLARE @new_id int;

 SET TRANSACTION ISOLATION LEVEL SERIALIZABLE;

 BEGIN TRY

 BEGIN TRANSACTION;

 SELECT @new_id = ISNULL(max(InvoiceNumber),0) + 1
 FROM Invoice
 WHERE BranchNumber = @BranchNumber ;

 INSERT INTO Invoice (BranchNumber, InvoiceNumber)
 VALUES (@BranchNumber, @new_id) ;

 COMMIT TRAN

 SELECT @new_id AS id

 END TRY
 BEGIN CATCH
 ROLLBACK TRAN;

 SELECT NULL AS id;

 ;THROW;

 END CATCH

 SET TRANSACTION ISOLATION LEVEL READ COMMITTED;

 RETURN @new_id
END
GO

The SERIALIZABLE isolation level is what makes this correct: it takes a range lock on the rows for the branch being numbered, so two concurrent sessions cannot read the same maximum and allocate the same number.

Data Modeling on Rudi Bruchez

Choosing a clustered index

When to choose a clustered index

Managing several auto-increments