Export (0) Print
Expand All

PipelineBuffer Class

Provides an in-memory data store containing rows and columns of data.

Namespace:  Microsoft.SqlServer.Dts.Pipeline
Assembly:  Microsoft.SqlServer.PipelineHost (in Microsoft.SqlServer.PipelineHost.dll)

public class PipelineBuffer : IDisposable

The PipelineBuffer is an in-memory two-dimensional data store containing rows and columns. It is created by the data flow task and supplied to managed data flow components during execution. The columns contained in a buffer are based on the columns in the IDTSOutputColumnCollection100 collections of the components in the graph.

Source components and components with asynchronous outputs receive a buffer for each of the output objects that are connected to a downstream component. These buffers are referred to as output buffers and do not contain rows. The component that receives the output buffer adds rows to the buffer and calls the SetEndOfRowset method when finished. This method sets the value of the EndOfRowset property to true on the final buffer. The data flow task then provides that buffer to the next component in the graph.

Transformation components with synchronous outputs and destination components receive PipelineBuffer objects in the ProcessInput method. The PipelineBuffer received in this method is an Input buffer and contains the rows that were added to by upstream components. This buffer is restricted and cannot be used to add or remove rows from the buffer.

The PipelineBuffer is written in managed code and supports managed data flow component developers by marshalling data between managed code and the underlying IDTSBuffer100 COM object.

For a complete list of Integration Services data types and the corresponding Get and Set methods of the PipelineBuffer class to use with each data type, see Working with Data Types in the Data Flow.

The following example shows a transformation component that iterates the rows and columns of a PipelineBuffer in ProcessInput.

using System;
using Microsoft.SqlServer.Dts;
using Microsoft.SqlServer.Dts.Pipeline;
using Microsoft.SqlServer.Dts.Pipeline.Wrapper;
using Microsoft.SqlServer.Dts.Runtime.Wrapper;


namespace Microsoft.Samples.SqlServer.Dts
{
   [DtsPipelineComponent
   (
   DisplayName="SampleComponent",
   ComponentType=ComponentType.Transform
   )]
   public class SampleComponent: PipelineComponent
   {
      public override void ProvideComponentProperties()
      {
         base.ProvideComponentProperties();

         ///Name the input and output add by the base class.
         ComponentMetaData.InputCollection[0].Name = "SampleInput";
         ComponentMetaData.OutputCollection[0].Name = "SampleOutput";
      }

      public override void ProcessInput(int inputID, PipelineBuffer buffer)
      {

      IDTSInput100 input = ComponentMetaData.InputCollection.GetObjectByID(inputID);

      while (buffer.NextRow())
      {
         foreach (IDTSInputColumn100 col in input.InputColumnCollection)
         {
            int colIndex = BufferManager.FindColumnByLineageID(input.Buffer,col.LineageID);
            object colData = buffer[colIndex];
            //TODO: Do something with the column data.
         }
      }
      }
   }
}

The following example shows a source component that adds rows to the output buffer in PrimeOutput.

using System;
using Microsoft.SqlServer.Dts;
using Microsoft.SqlServer.Dts.Pipeline;
using Microsoft.SqlServer.Dts.Pipeline.Wrapper;
using Microsoft.SqlServer.Dts.Runtime.Wrapper;


namespace Microsoft.Samples.SqlServer.Dts
{
   [DtsPipelineComponent
   (
      DisplayName="SampleComponent",
      ComponentType=ComponentType.SourceComponent
)]
   public class SampleComponent: PipelineComponent
   {
      public override void PrimeOutput(int outputs, int[] outputIDs,PipelineBuffer[] buffers)
      {
         int rows = 100;
         PipelineBuffer buf = buffers[0];
         IDTSOutput100 output = ComponentMetaData.OutputCollection[0];
         Random rand = new Random();

         //Loop rows number of times
         for(int r = 0; r < rows; r++)
         {
            buf.AddRow();
            foreach( IDTSOutputColumn100 col in output.OutputColumnCollection)
            {
               int colIndex = BufferManager.FindColumnByLineageID( output.Buffer, col.LineageID);
               // Note, buffer columns containing binary large objects
               // can not be set using the following syntax. Instead,
               // the AddBlobData and SetBytes methods are used.
               buf[colIndex] = rand.Next();
            }
         }
         buf.SetEndOfRowset();
      }
   }
}

System.Object
  Microsoft.SqlServer.Dts.Pipeline.PipelineBuffer

Any public static (Shared in Visual Basic) members of this type are thread safe. Any instance members are not guaranteed to be thread safe.
Was this page helpful?
(1500 characters remaining)
Thank you for your feedback

Community Additions

ADD
Show:
© 2014 Microsoft