Why should you care?

Range scans on OrderDate or sequential Id lookups are fastest when the clustered key matches how you access data.

See it live — copy this example

Run in SQL Server Management Studio (SSMS) or Azure Data Studio.

USE DataVerse;
IF OBJECT_ID(N'dbo.Invoices', N'U') IS NOT NULL DROP TABLE dbo.Invoices;
CREATE TABLE dbo.Invoices (
    InvoiceId INT NOT NULL,
    CustomerId INT NOT NULL,
    InvoiceDate DATE NOT NULL,
    Total DECIMAL(12,2) NOT NULL,
    CONSTRAINT PK_Invoices PRIMARY KEY CLUSTERED (InvoiceId)
);
CREATE UNIQUE NONCLUSTERED INDEX UX_Invoices_BizKey
    ON dbo.Invoices (CustomerId, InvoiceDate, InvoiceId);
SELECT i.name, i.type_desc
FROM sys.indexes i
WHERE i.object_id = OBJECT_ID(N'dbo.Invoices');

What happened?

• PRIMARY KEY CLUSTERED stores rows in InvoiceId order.
• A separate nonclustered unique index supports business lookups without changing the clustered choice.

Practice next

1. Create Invoices and list indexes.
2. Insert a few rows and SELECT WHERE InvoiceId = …
3. View the estimated plan — should seek on clustered.
4. Rebuild: ALTER INDEX PK_Invoices ON dbo.Invoices REBUILD;
5. Compare plans for WHERE CustomerId = @id using the NC index.

Remember

Clustered index = row order. Usually the primary key. Choose the key for common access patterns.

Invoice Id lookups

Billing API loads DataVerse invoices by InvoiceId constantly.

Outcome: Clustered PK keeps those gets cheap.