This documentation is archived and is not being maintained.

Performing Batch Operations Using DataAdapters (ADO.NET)

Batch support in ADO.NET allows a DataAdapter to group INSERT, UPDATE, and DELETE operations from a DataSet or DataTable to the server, instead of sending one operation at a time. The reduction in the number of round trips to the server typically results in significant performance gains. Batch updates are supported for the .NET data providers for SQL Server (System.Data.SqlClient) and Oracle (System.Data.OracleClient).

When updating a database with changes from a DataSet in previous versions of ADO.NET, the Update method of a DataAdapter performed updates to the database one row at a time. As it iterated through the rows in the specified DataTable, it examined each DataRow to see if it had been modified. If the row had been modified, it called the appropriate UpdateCommand, InsertCommand, or DeleteCommand, depending on the value of the RowState property for that row. Every row update involved a network round-trip to the database.

Starting with ADO.NET 2.0, the DbDataAdapter exposes an UpdateBatchSize property. Setting the UpdateBatchSize to a positive integer value causes updates to the database to be sent as batches of the specified size. For example, setting the UpdateBatchSize to 10 will group 10 separate statements and submit them as single batch. Setting the UpdateBatchSize to 0 will cause the DataAdapter to use the largest batch size that the server can handle. Setting it to 1 disables batch updates, as rows are sent one at a time.

Executing an extremely large batch could decrease performance. Therefore, you should test for the optimum batch size setting before implementing your application.

When batch updates are enabled, the UpdatedRowSource property value of the DataAdapter's UpdateCommand, InsertCommand, and DeleteCommand should be set to None or OutputParameters. When performing a batch update, the command's UpdatedRowSource property value of FirstReturnedRecord or Both is invalid.

The following procedure demonstrates the use of the UpdateBatchSize property. The procedure takes two arguments, a DataSet object that has columns representing the ProductCategoryID and Name fields in the Production.ProductCategory table, and an integer representing the batch size (the number of rows in the batch). The code creates a new SqlDataAdapter object, setting its UpdateCommand, InsertCommand, and DeleteCommand properties. The code assumes that the DataSet object has modified rows. It sets the UpdateBatchSize property and executes the update.

public static void BatchUpdate(DataTable dataTable,Int32 batchSize)
    // Assumes GetConnectionString() returns a valid connection string.
    string connectionString = GetConnectionString();

    // Connect to the AdventureWorks database.
    using (SqlConnection connection = new 

        // Create a SqlDataAdapter.
        SqlDataAdapter adapter = new SqlDataAdapter();

        // Set the UPDATE command and parameters.
        adapter.UpdateCommand = new SqlCommand(
            "UPDATE Production.ProductCategory SET "
            + "Name=@Name WHERE ProductCategoryID=@ProdCatID;", 
           SqlDbType.NVarChar, 50, "Name");
           SqlDbType.Int, 4, "ProductCategoryID");
         adapter.UpdateCommand.UpdatedRowSource = UpdateRowSource.None;

        // Set the INSERT command and parameter.
        adapter.InsertCommand = new SqlCommand(
            "INSERT INTO Production.ProductCategory (Name) VALUES (@Name);", 
          SqlDbType.NVarChar, 50, "Name");
        adapter.InsertCommand.UpdatedRowSource = UpdateRowSource.None;

        // Set the DELETE command and parameter.
        adapter.DeleteCommand = new SqlCommand(
            "DELETE FROM Production.ProductCategory "
            + "WHERE ProductCategoryID=@ProdCatID;", connection);
          SqlDbType.Int, 4, "ProductCategoryID");
        adapter.DeleteCommand.UpdatedRowSource = UpdateRowSource.None;

        // Set the batch size.
        adapter.UpdateBatchSize = batchSize;

        // Execute the update.

The DataAdapter has two update-related events: RowUpdating and RowUpdated. In previous versions of ADO.NET, when batch processing is disabled, each of these events is generated once for each row processed. RowUpdating is generated before the update occurs, and RowUpdated is generated after the database update has been completed.

Event Behavior Changes with Batch Updates

When batch processing is enabled, multiple rows are updated in a single database operation. Therefore, only one RowUpdated event occurs for each batch, whereas the RowUpdating event occurs for each row processed. When batch processing is disabled, the two events are fired with one-to-one interleaving, where one RowUpdating event and one RowUpdated event fire for a row, and then one RowUpdating and one RowUpdated event fire for the next row, until all of the rows are processed.

Accessing Updated Rows

When batch processing is disabled, the row being updated can be accessed using the Row property of the RowUpdatedEventArgs class.

When batch processing is enabled, a single RowUpdated event is generated for multiple rows. Therefore, the value of the Row property for each row is null. RowUpdating events are still generated for each row. The CopyToRows method of the RowUpdatedEventArgs class allows you to access the processed rows by copying references to the rows into an array. If no rows are being processed, CopyToRows throws an ArgumentNullException. Use the RowCount property to return the number of rows processed before calling the CopyToRows method.

Handling Data Errors

Batch execution has the same effect as the execution of each individual statement. Statements are executed in the order that the statements were added to the batch. Errors are handled the same way in batch mode as they are when batch mode is disabled. Each row is processed separately. Only rows that have been successfully processed in the database will be updated in the corresponding DataRow within the DataTable.

The data provider and the back-end database server determine which SQL constructs are supported for batch execution. An exception may be thrown if a non-supported statement is submitted for execution.