SQL CREATE TABLE Generator

How the SQL CREATE TABLE Generator works

Every change you make is turned into a CREATE TABLE statement straight away. Names are quoted for the chosen dialect ([brackets], "double quotes" or `backticks`), or only when needed if you untick Quote all names. The type list shows the types of the selected database, and the length or precision field is checked against the type: NVARCHAR needs a length or MAX, DECIMAL takes precision and scale such as 10,2, and types like INT take none.

Identity columns are written as IDENTITY(1, 1) for SQL Server, GENERATED ALWAYS AS IDENTITY for PostgreSQL and AUTO_INCREMENT for MySQL. Primary key and identity columns are always NOT NULL. One primary key column is marked inline; several are combined into a CONSTRAINT PK_table PRIMARY KEY (...) line. On SQL Server, columns that allow nulls are written with an explicit NULL. Problems such as a missing length or two identity columns are explained in the status line. The statement is not run against any database.

How to use the SQL CREATE TABLE Generator

1. Choose the dialect, then enter the table name and, if you use one, the schema (for example dbo or public).
2. For each column enter a name, pick a type, and fill in the length or precision where the type needs one, such as 255 for NVARCHAR or 10,2 for DECIMAL.
3. Tick Primary key, NOT NULL, Identity (Auto increment in MySQL) or Unique as needed, and enter a default value if the column has one.
4. Use Add column, Up, Down and Remove to change the column list, or select Load example to start from a sample table.
5. Copy the CREATE TABLE statement or download it as a .sql file. Any problem, such as a missing length, is explained just above the output.

Example

Table Customers in schema dbo on SQL Server, with CustomerId INT (primary key, identity), Email NVARCHAR 255 (NOT NULL, unique), CreditLimit DECIMAL 10,2 (NOT NULL, default 0) and Status VARCHAR 20 (NOT NULL, default active):

CREATE TABLE [dbo].[Customers] (
    [CustomerId] INT IDENTITY(1, 1) NOT NULL PRIMARY KEY,
    [Email] NVARCHAR(255) NOT NULL UNIQUE,
    [CreditLimit] DECIMAL(10, 2) NOT NULL DEFAULT 0,
    [Status] VARCHAR(20) NOT NULL DEFAULT 'active'
);

Switching the dialect to PostgreSQL with schema public changes it to:

CREATE TABLE "public"."Customers" (
    "CustomerId" INTEGER GENERATED ALWAYS AS IDENTITY NOT NULL PRIMARY KEY,
    "Email" VARCHAR(255) NOT NULL UNIQUE,
    "CreditLimit" NUMERIC(10, 2) NOT NULL DEFAULT 0,
    "Status" VARCHAR(20) NOT NULL DEFAULT 'active'
);

Common use cases

• Sketching a new table during design without remembering the exact syntax for each database.
• Writing the same table for SQL Server, PostgreSQL and MySQL by switching the dialect; column types are converted to the closest equivalent.
• Creating a table for test data before generating INSERT statements.
• Learning how identity columns, composite primary keys and defaults are written in each dialect.

Common errors and how to fix them

Enter a table name / Add at least one column

The generator needs both before it can produce a statement. Fill in the table name and at least one column definition.

Only one column per table can be an identity / auto increment column

You have marked more than one column as auto-incrementing, which no supported database allows. Keep the flag on the single surrogate key column.

Unexpected text after a quoted part of the table name

The table name mixes quoted and unquoted parts, for example a stray bracket or bracketed schema. Enter the plain name, or a schema-qualified name such as sales.orders, and let the tool apply the dialect's quoting.

A column name is rejected as a reserved word

The name collides with a SQL keyword. Rename it, or rely on the dialect's identifier quoting, keeping in mind that quoted names must then be quoted everywhere they are used.

Frequently asked questions