Example - Restarting A Failed Process From A Checkpoint File; Example - Restarting A Failed Process From The Beginning; Reducing Build Times For Large Powercubes - IBM Cognos User Manual

Procedure
The following MDL file contains the verb statements that are required to generate
categories for the model go_sales.py?, without creating the cubes for that model.
OpenPY "go_sales.py?" PopulateModel SavePY "go_sales.py?"
If these statements are saved in a file called Gen_nat.mdl, you can process them by
running Cognos Transformer, as follows:
cogtr -mGen_nat.mdl

Example - Restarting a Failed Process from a Checkpoint File

As Cognos Transformer processes data on the server, it maintains a checkpoint file
with the extension .qy?. When a cube build or update fails, Cognos Transformer
can use this .qy? file to restart processing at the point of failure.
For example, suppose you are running a quarterly model update with new data.
Cognos Transformer is unable to locate one of the source files for a data source in
the model, and the model update fails. You use the checkpoint file to restart
processing at the point of failure.
You can use the following command to resume processing of the .py? model file:
cogtr -i go_sales.py?
If an associated .qy? file is found, it is used to restart the process.
For more information, see Appendix A, "Command Line Options," on page 209.

Example - Restarting a Failed Process from the Beginning

When Cognos Transformer builds cubes, some conditions may result in a failed
process.
For example, if Cognos Transformer has insufficient disk space for the model files,
data temporary files, and cubes, an error occurs and processing stops. In such
cases, you can restart the process from the beginning, using the following
command:
cogtr -i go_sales.py?
This command loads the saved model file, go_sales.py?, ignoring the last
interrupted cube creation process and any associated .qy? files.
For more information, see Appendix A, "Command Line Options," on page 209.

Reducing Build Times for Large PowerCubes

This section describes techniques you can use to reduce build times when working
with cubes having more than 500,000 categories, and parent:child ratios that
deviate significantly from the ideal.
Included are examples to illustrate the efficiency gains you can expect from
optimization strategies such as hardware upgrades, system tuning, memory
re-allocation, or model redesign.
The main test model used to confirm these strategies had the characteristics
described in the following table.
202
IBM Cognos Transformer Version 10.1.1: User Guide