Stored Procedures (Why Use Stored Procedures? )

In this article, we will examine what a stored procedure is and how to call one from SQL Server database.

Outline:

• Why Use Stored Procedures?
• Creating a Stored Procedure
• Calling a Stored Procedure
• Specifying Parameters
• Data Retrieval
• Inserting Data Using Parameters

Why Use Stored Procedures?

There are several advantages of using stored procedures instead of standard SQL. First, stored procedures allow a lot more flexibility offering capabilities such as conditional logic. Second, because stored procedures are stored within the DBMS, bandwidth and execution time are reduced. This is because a single stored procedure can execute a complex set of SQL statements. Third, SQL Server pre-compiles stored procedures such that they execute optimally. Fourth, client developers are abstracted from complex designs. They would simply need to know the stored procedure's name and the type of data it returns.

Creating a Stored Procedure

Enterprise Manager provides an easy way to create stored procedures. First, select the database to create the stored procedure on. Expand the database node, right-click on "Stored Procedures" and select "New Stored Procedure...". You should see the following:

CREATE PROCEDURE [OWNER].[PROCEDURE NAME] AS

Substitute OWNER with "dbo" (database owner) and PROCEDURE NAME with the name of the procedure. For example:

CREATE PROCEDURE [dbo].[GetProducts] AS

So far, we are telling SQL Server to create a new stored procedure with the name GetProducts. We specify the body of the procedure after the AS clause:

CREATE PROCEDURE [dbo].[GetProducts] AS SELECT ProductID, ProductName FROM Products

Click on the Check Syntax button in order to confirm that the stored procedure is syntactically correct. Please note that the GetProducts example above will work on the Northwind sample database that comes with SQL Server. Modify it as necessary to suite the database you are using.

Now that we have created a stored procedure, we will examine how to call it from within a C# application.

Calling a Stored Procedure

A very nice aspect of ADO.NET is that it allows the developer to call a stored procedure in almost the exact same way as a standard SQL statement.

1. Create a new C# Windows Application project.

2. From the Toolbox, drag and drop a DataGrid onto the Form. Resize it as necessary.

3. Double-click on the Form to generate the Form_Load event handler. Before entering any code, add "using System.Data.SqlClient" at the top of the file.

Enter the following code:

private void Form1_Load(object sender, System.EventArgs e) { SqlConnection conn = new SqlConnection("Data Source=localhost;Database=Northwind;Integrated Security=SSPI"); SqlCommand command = new SqlCommand("GetProducts", conn); SqlDataAdapter adapter = new SqlDataAdapter(command); DataSet ds = new DataSet(); adapter.Fill(ds, "Products"); this.dataGrid1.DataSource = ds; this.dataGrid1.DataMember = "Products"; }

As you can see, calling a stored procedure in this example is exactly like how you would use SQL statements, only that instead of specifying the SQL statement, you specify the name of the stored procedure. Aside from that, you can treat it exactly the same as you would an ordinary SQL statement call with all the advantages of a stored procedure.

Specifying Parameters

Most of the time, especially when using non-queries, values must be supplied to the stored procedure at runtime. For instance, a @CategoryID parameter can be added to our GetProducts procedure in order to specify to retrieve only products of a certain category. In SQL Server, parameters are specified after the procedure name and before the AS clause.

CREATE PROCEDURE [dbo].[GetProducts] (@CategoryID int) AS SELECT ProductID, ProductName FROM Products WHERE CategoryID = @CategoryID

Parameters are enclosed within parenthesis with the parameter name first followed by the data type. If more than one parameter is accepted, they are separated by commas:

CREATE PROCEDURE [dbo].[SomeProcedure] ( @Param1 int, @Param2 varchar(50), @Param3 varchar(50) ) AS ...

For our GetProducts example, if @CategoryID was supplied with the value 1, the query would equate to:

SELECT ProductID, ProductName FROM Products WHERE CategoryID = 1

Which would select all the products that belong to CategoryID 1 or the Beverages category. To call the stored procedure, use Query Analyzer to execute:

exec GetProducts X

where X is the @CategoryID parameter passed to the stored procedure. To call the stored procedure from within a C# application using 1 as the @CategoryID parameter value, use the following code:

SqlConnection conn = new SqlConnection("Data Source=localhost;Database=Northwind;Integrated Security=SSPI"); SqlCommand command = new SqlCommand("GetProducts", conn); command.CommandType = CommandType.StoredProcedure; command.Parameters.Add("@CategoryID", SqlDbType.Int).Value = 1; SqlDataAdapter adapter = new SqlDataAdapter(command); DataSet ds = new DataSet(); adapter.Fill(ds, "Products"); this.dataGrid1.DataSource = ds; this.dataGrid1.DataMember = "Products";

Note that you must now specify the CommandType property of the SqlCommand object. The reason we did not do this in the first example was that it is not required if the stored procedure does not accept parameters. Of course, specifying the CommandType property even if it is not needed may improve readability. The next line actually combines two lines in one:

command.Parameters.Add("@CategoryID", SqlDbType.Int); command.Parameters["@CategoryID"].Value = 1;

The first line of this segment specifies that the command object (which calls the GetProducts stored procedure) accepts a parameter named @CategoryID which is of type SqlDbType.Int. The type must be the same as the data type specified by the stored procedure. The second line of this code segment gives the parameter the value 1. For simplicity, especially when using more than one parameter, I prefer to combine to two lines into a single line:

command.Parameters.Add("@CategoryID", SqlDbType.Int).Value = 1;

The rest of the code is the same as in the previous example without parameters. As illustrated in the previous examples, ADO.NET takes a lot of pain out of database programming. Calling a stored procedure uses virtually the same code as using standard SQL and specifying parameters is a painless process.

Data Retrieval

Data Retrieval with stored procedures is the same (surprise!) as if using standard SQL. You can wrap a DataAdapter around the Command object or you can use a DataReader to fetch the data one row at a time. The previous examples have already illustrated how to use a DataAdapter and fill a DataSet. The following example shows usage of the DataReader:

SqlConnection conn = new SqlConnection("Data Source=localhost;Database=Northwind;Integrated Security=SSPI"); SqlCommand command = new SqlCommand("GetProducts", conn); command.CommandType = CommandType.StoredProcedure; command.Parameters.Add("@CategoryID", SqlDbType.Int).Value = 1; conn.Open(); SqlDataReader reader = command.ExecuteReader(); while (reader.Read()) { Console.WriteLine(reader["ProductName"]); } conn.Close();

Again, using either a DataAdapter or a DataReader against a query from a stored procedure is the same as specifying the SQL from within the code.

Inserting Data Using Parameters

Using other SQL statements such as INSERT, UPDATE or DELETE follow the same procedure. First, create a stored procedure that may or may not accept parameters, and then call the stored procedure from within the code supply the necessary values if parameters are needed. The following example illustrates how to insert a new user in a users table that has a username and password field.

CREATE PROCEDURE [dbo].[InsertUser] ( @Username varchar(50), @Password varchar(50) ) AS INSERT INTO Users VALUES(@Username, @Password)

string username = ... // get username from user string password = ... // get password from user SqlConnection conn = new SqlConnection("Data Source=localhost;Database=MyDB;Integrated Security=SSPI"); SqlCommand command = new SqlCommand("InsertUser", conn); command.CommandType = CommandType.StoredProcedure; command.Parameters.Add("@Username", SqlDbType.VarChar).Value = username; command.Parameters.Add("@Password", SqlDbType.VarChar).Value = password; conn.Open(); int rows = command.ExecuteNonQuery(); conn.Close();

First, we retrieve the username and password information from the user. This information may be entered onto a form, through a message dialog or through some other method. The point is, the user specifies the username and password and the applicaton inserts the data into the database. Also notice that we called the ExecuteNonQuery() method of the Connection object. We call this method to indicate that the stored procedure does not return results for a query but rather an integer indicating how many rows were affected by the executed statement. ExecuteNonQuery() is used for DML statements such as INSERT, UPDATE and DELETE. Note that we can test the value of rows to check if the stored procedure inserted the data successfully.

if (rows == 1) { MessageBox.Show("Create new user SUCCESS!"); } else { MessageBox.Show("Create new user FAILED!"); }

We check the value of rows to see if it is equal to one. Since our stored procedure only did one insert operation and if it is successful, the ExecuteNonQuery() method should return 1 to indicate the one row that was inserted. For other SQL statements, especially UPDATE and DELETE statements that affect more than one row, the stored procedure will return the number of rows affected by the statement.

DELETE FROM Products WHERE ProductID > 50

This will delete all products whose product ID is greater than 50 and will return the number of rows deleted.