About the Talk
November 4, 2010 9:00 AM
Date/Time TBA Microsoft designed Integration Services with ETL processing for Data Warehouses in mind, so loading one should be pretty easy. It's not that bad, but there are a few things you should avoid and non-obvious best practices to use to make that ETL faster and more reliable. Assuming attendees have a basic level of familiarity with Kimball Data Warehousing methodology, this session will explain how to accomplish Kimball style ETL using SQL Server Integration Services. We'll review the SCD Wizard and how it works for processing your dimension tables - what it's good for, what it's not, and what alternatives you have to it. We'll also examine fact table loading using a surrogate key replacement pipeline, especially using the Lookup component efficiently. Finally, we’ll look at specific and easy performance tweaks, techniques and best practices that apply directly to DW loads. This includes exposing which Data Flow components to avoid, how to still get your work done without them, as well as tuning inserts and updates.