SQL Integer Types Compared

INT, BIGINT, SMALLINT, and TINYINT

All four store whole numbers without decimal places. The main differences are their range and storage size.

In SQL Server, TINYINT does not support negative values.

Examples and uses

CREATE TABLE Employees
(
    EmployeeId  INT,       -- Normal identifier
    Age         TINYINT,   -- 0 to 255
    DepartmentId SMALLINT, -- Small range of IDs
    TotalViews  BIGINT     -- Very large count
);

TINYINT

Suitable for very small non-negative numbers:

Age TINYINT
Rating TINYINT
MonthNumber TINYINT

SMALLINT

Suitable when the value can exceed 255 but remains relatively small:

YearNumber SMALLINT
Quantity SMALLINT

INT

The most commonly used integer type:

EmployeeId INT
OrderId    INT
Quantity   INT

It is generally the default choice unless the business range requires a smaller or larger type.

BIGINT

Suitable for extremely large values:

TransactionId BIGINT
TotalPageViews BIGINT
LargeSequenceId BIGINT

In C#, SQL Server BIGINT maps to long.

Why does the correct size matter?

Choosing the smallest suitable type can reduce:

• Table storage

• Index size

• Memory usage

• Disk reads

• Network transfer

However, do not choose a smaller type if the value may exceed its range. Changing a primary key from INT to BIGINT later can be difficult because related foreign keys and indexes must also be changed.

Identity example

CREATE TABLE Orders
(
    OrderId BIGINT IDENTITY(1,1) PRIMARY KEY,
    Quantity SMALLINT
);

IDENTITY automatically generates values, but the column is still limited by its chosen data type.

Key interview points

• All four types store exact whole numbers.

• Their main differences are range and storage size.

• TINYINT is unsigned in SQL Server: 0–255.

• INT is the usual general-purpose choice.

• Use BIGINT only when values may exceed the INT range.

• Primary and foreign-key columns should use the same data type.

• Larger key types increase table and index storage.

• Integer overflow occurs when a value exceeds the type’s supported range.

• They do not store decimal values; use DECIMAL when a fractional part is required.