Prepare sample data

There are several ways to prepare sample data to be used for generated business objects in a benchmark:

Using Test Connector page "Using Test Connector"
Using Map Designer page "Using Map Designer"
Using tracing output page "Using tracing output"
Generating workload to a file page "Generating workload to a file"

Using Test Connector

You can save a business object to a file from Test Connector and use it as an input file of sample data for a benchmark.

To implement this approach, see Chapter 9, Test Connector Tool in the Collaboration Development Guide.

The advantages of this approach are:

The disadvantages of this approach are:

Using Map Designer

You can save a business object to a file from Map Designer and use it as an input file of sample data for a benchmark.

To implement this approach, see Chapter 4, Compiling and Testing Maps in the Map Development Guide.

The advantages of this approach are:

The disadvantage of this approach is that you must create all the test data by hand. If you use the technique "Generating workload to a file", then the process is automated.

Using tracing output

You can copy business object data that is output to the tracing destination of a connector at higher trace levels and use it as an input file of sample data for a benchmark.

To implement this approach, do the following:

  1. Set the trace level for the connector to a level at which it outputs the contents of the business objects it processes.
  2. Trigger a business object of the appropriate type and have the connector process it.
  3. Copy the business object contents from the tracing destination into a file.

    Business object data is recorded in a kind of markup language: elements like attributes, child objects, and the entire top-level business object are demarcated by pairs of tags. Each business object begins with a set of StartHeader and EndHeader tags, which is followed by a StartBO tag that includes the type of the top-level business object. Below is an example:

    <StartHeader>
    <Version = 3.0>
    <EndHeader>
    <StartBO:BusinessObjectType>
    

    The output for a business object ends with an EndBO tag that includes the type of the top-level business object just as the StartBO tag does. Below is an example:

    <EndBO:BusinessObjectType>
    

    Copy all of the output from the StartHeader tag to the EndBO tag.

    Important:
    The StartHeader tag might appear on a line that also contains other text (such as information about the tracing subsystem), and the EndBO tag typically appears on a line that ends with a right bracket (]). Be sure to include only the important tags; do not copy unnecessary characters, or the system cannot use the file as input.
  4. Save the file.

The advantages of this approach are:

The disadvantages of this approach are:

Generating workload to a file

You can use a feature of the benchmark wizard to generate a file of sample data.

To implement this approach, do the following:

  1. Determine the location and name that you want for the file of sample data to be generated.
    Note:
    If you plan ahead this way then you only have to run the benchmark wizard two times--one time to create the benchmark definition and a second time to generate the input file. If you do not plan ahead then you have to run the benchmark wizard a third time to reconfigure the benchmark definition to use the input file.
  2. Prepare a benchmark definition. Follow the instructions in the section "Preparing a benchmark definition".
    1. When you are presented with the Benchmark Components screen, specify the file path and name you determined in step 1, as described in step 7 of the section "Configuring benchmark components".
    2. Proceed through the Object Properties screen and finish the benchmark definition.
  3. Follow the instructions in the section "Generating workload to a file" to create a file of sample data for the benchmark definition created in step 2.

The advantage of this approach is: that you can conveniently generate a large amount of sample data through configuration and metadata.

The disadvantages of this approach are:

Copyright IBM Corp. 1997, 2004