1. In SQL Server Management Studio, click New Query to display the Query Editor.

2. Copy the Transact-SQL code from the following example into the Query Editor. This Transact-SQL code creates a FILESTREAM-enabled database called Archive.

   Note:
   For this script, the directory C:\Data must exist.

3. To build the database, click Execute.

The following code example creates a database that is named Archive. The database contains three filegroups: PRIMARY, Arch1, and FileStreamGroup1. PRIMARY and Arch1 are regular filegroups that cannot contain FILESTREAM data. FileStreamGroup1 is the FILESTREAM filegroup.

CREATE DATABASE Archive 
ON
PRIMARY ( NAME = Arch1,
    FILENAME = 'c:\data\archdat1.mdf'),
FILEGROUP FileStreamGroup1 CONTAINS FILESTREAM( NAME = Arch3,
    FILENAME = 'c:\data\filestream1')
LOG ON  ( NAME = Archlog1,
    FILENAME = 'c:\data\archlog1.ldf')
GO

For a FILESTREAM filegroup, FILENAME refers to a path. The path up to the last folder must exist, and the last folder must not exist. In this example, c:\data must exist. However, the filestream1 subfolder cannot exist when you execute the CREATE DATABASE statement. For more information about the syntax, see CREATE DATABASE (Transact-SQL).

After you run the previous example, a filestream.hdr file and an $FSLOG folder appears in the c:\Data\filestream1 folder. The filestream.hdr file is a header file for the FILESTREAM container.

Important:
The filestream.hdr file is an important system file. It contains FILESTREAM header information. Do not remove or modify this file.

For existing databases, you can use the ALTER DATABASE statement to add a FILESTREAM filegroup.

Concepts

Other Resources

Help and Information