Batch Processing
The CData ADO.NET Provider for Apache Phoenix enables you to take advantage of the bulk load support in Apache Phoenix through ApachePhoenixDataAdapters. You can use the Batch API to execute related SQL data manipulation statements simultaneously.
Using the ADO.NET Batch API
Performing a batch update consists of the following basic steps:
- Define custom parameterized SQL statements in ApachePhoenixCommand objects.
- Set the UpdatedRowSource property of the ApachePhoenixCommand object to "UpdateRowSource.None".
- Assign the ApachePhoenixCommand objects to the ApachePhoenixDataAdapter.
- Add the parameters to the command.
- Call the ApachePhoenixDataAdapter's Update method. Pass in a DataSet or DataTable containing your changes.
Controlling Batch Size
Depending on factors such as the size of the request, your network resources, and the performance of the server, you may gain performance by executing several smaller batch requests. You can control the size of each batch by setting the ApachePhoenixDataAdapter's UpdateBatchSize property to a positive integer.
Bulk INSERT
The following code prepares a single batch that inserts records in bulk and retrieves the new records' Ids. The example executes a batch INSERT of new DataRows, which have the "Added" state.
C#
ApachePhoenixDataAdapter adapter = new ApachePhoenixDataAdapter(); using (ApachePhoenixConnection conn = new ApachePhoenixConnection("URL=http://localhost:8765;")) { conn.Open(); adapter.InsertCommand = conn.CreateCommand(); adapter.InsertCommand.CommandText = "INSERT INTO [SampleCatalog_1].[SampleSchema_1].SampleTable_1 (Name) VALUES (@Name)"; adapter.InsertCommand.UpdatedRowSource = UpdateRowSource.None; adapter.InsertCommand.Parameters.Add("@Name", "Name"); DataTable batchDataTable = new DataTable(); batchDataTable.Columns.Add("Name", typeof(string)); batchDataTable.Rows.Add("Jon Doe"); batchDataTable.Rows.Add("John"); adapter.UpdateBatchSize = 2; adapter.Update(batchDataTable); ApachePhoenixCommand cmd = new ApachePhoenixCommand("SELECT * FROM LastResultInfo#TEMP", conn); adapter = new ApachePhoenixDataAdapter(cmd); DataTable res = new DataTable(); adapter.Fill(res); foreach (DataRow row in res.Rows) foreach(DataColumn col in res.Columns) Console.Write("{0}: {1}", col.ColumnName, row[col]); }
VB.NET
Dim adapter As New ApachePhoenixDataAdapter()
Using conn As New ApachePhoenixConnection("URL=http://localhost:8765;")
conn.Open()
adapter.InsertCommand = conn.CreateCommand()
adapter.InsertCommand.CommandText = "INSERT INTO [SampleCatalog_1].[SampleSchema_1].SampleTable_1 (Id) VALUES (@Name)"
adapter.InsertCommand.UpdatedRowSource = UpdateRowSource.None
adapter.InsertCommand.Parameters.Add("@Name", "Name")
Dim batchDataTable As New DataTable()
batchDataTable.Columns.Add("Name", GetType(String))
batchDataTable.Rows.Add("John")
batchDataTable.Rows.Add("Jon Doe")
adapter.UpdateBatchSize = 2
adapter.Update(batchDataTable)
Dim cmd As New ApachePhoenixCommand("SELECT * FROM LastResultInfo#TEMP", conn)
adapter = New ApachePhoenixDataAdapter(cmd)
Dim res As New DataTable()
adapter.Fill(res)
For Each row As DataRow In res.Rows
For Each col As DataColumn In res.Columns
Console.WriteLine("{0}: {1}", col.ColumnName, row(col))
Next
Next
End Using
Bulk Update
A batch update additionally requires the primary key of each row to update. The following example executes a batch for all DataRow records with a "Modified" state:
C#
ApachePhoenixDataAdapter adapter = new ApachePhoenixDataAdapter(); using (ApachePhoenixConnection conn = new ApachePhoenixConnection("URL=http://localhost:8765;")) { conn.Open(); adapter.UpdateCommand = conn.CreateCommand(); adapter.UpdateCommand.CommandText = "UPDATE [SampleCatalog_1].[SampleSchema_1].SampleTable_1 SET Name=@Name WHERE Id=@Id"; adapter.UpdateCommand.Parameters.Add("@Name", "Name"); adapter.UpdateCommand.Parameters.Add("@Id", "Id"); adapter.UpdateCommand.UpdatedRowSource = UpdateRowSource.None; DataTable batchDataTable = new DataTable(); batchDataTable.Columns.Add("Name", typeof(string)); batchDataTable.Rows.Add("Jon Doe"); batchDataTable.Rows.Add("John"); adapter.UpdateBatchSize = 2; adapter.Update(dataTable); }
VB.NET
Dim adapter As New ApachePhoenixDataAdapter()
Using conn As New ApachePhoenixConnection("URL=http://localhost:8765;")
conn.Open()
adapter.UpdateCommand = conn.CreateCommand()
adapter.UpdateCommand.CommandText = "UPDATE [SampleCatalog_1].[SampleSchema_1].SampleTable_1 SET Name=@Name WHERE Id=@Id"
adapter.UpdateCommand.Parameters.Add("@Name", "Name")
adapter.UpdateCommand.Parameters.Add("@Id", "Id")
adapter.UpdateCommand.UpdatedRowSource = UpdateRowSource.None
Dim batchDataTable As New DataTable()
batchDataTable.Columns.Add("Name", GetType(String))
batchDataTable.Rows.Add("John")
batchDataTable.Rows.Add("Jon Doe")
adapter.UpdateBatchSize = 2
adapter.Update(dataTable)
End Using
Bulk Delete
The following code prepares a single batch that deletes records in bulk. The primary key for each row is required. The following example executes a batch for all DataRow records with a "Deleted" state:
C#
ApachePhoenixDataAdapter adapter = new ApachePhoenixDataAdapter();
using (ApachePhoenixConnection conn = new ApachePhoenixConnection("URL=http://localhost:8765;")) {
conn.Open();
adapter.DeleteCommand = conn.CreateCommand();
adapter.DeleteCommand.CommandText = "DELETE FROM [SampleCatalog_1].[SampleSchema_1].SampleTable_1 WHERE Id=@Id";
adapter.DeleteCommand.Parameters.Add("@Id", "Id");
adapter.DeleteCommand.UpdatedRowSource = UpdateRowSource.None;
adapter.UpdateBatchSize = 2;
adpater.Update(table);
}
VB.NET
Dim adapter As New ApachePhoenixDataAdapter()
Using conn As New ApachePhoenixConnection("URL=http://localhost:8765;")
conn.Open()
adapter.DeleteCommand = conn.CreateCommand()
adapter.DeleteCommand.CommandText = "DELETE FROM [SampleCatalog_1].[SampleSchema_1].SampleTable_1 WHERE Id=@Id"
adapter.DeleteCommand.Parameters.Add("@Id", "Id")
adapter.DeleteCommand.UpdatedRowSource = UpdateRowSource.None
adapter.UpdateBatchSize = 2
adpater.Update(table)
End Using