Simplify Mappings.

Keep your assignments as simple as possible.

Smaller is better in terms of performance and tuning.

Divide and conquer is the best strategy for achieving the fastest mapping performance.

In some cases, multiple phases of work, or splitting the workload between databases and phases, can resolve upstream dependencies and improve concurrency.

Mappings containing more than 50 objects are too large and must be split into multiple mappings.

When replacing PERL code, be sure to break the code into units of work.

Use each unit as a design step in your mapping architecture.

Create an entire complex mapping and break it down into smaller manageable steps.

Streamline Mapping:

Make your mapping objects as reasonable as possible.

Run your data through transformations rather than around transformations.

This is useful for session-level partitioning options and mapping concurrency capabilities.

Use Variable Ports for complex expressions:

Break down complex problems into smaller parts.

Use variables in expressions to create complex expression logic.

This makes the mapping more maintainable.

Minimize the number of targets:

Don't set more than 5 goals per assignment.

This slows down allocation dramatically.

Complex maps typically require multiple goals, and performance decreases as the number of goals increases.

Use 64-bit PowerCenter when processing large amounts of data:

If you are running very large and complex mappings with large amounts of data (approximately 50 million rows or more), we recommend using PowerCenter 64-bit servers for optimal performance.

This ensures that even large mappings have access to sufficient storage and fast performance.