Identifying High Risk SQL Before The Upgrade

• Ensure a smooth Oracle upgrade for your company •

Everyone knows identifying the high risk Oracle database SQL AFTER the upgrade is easy. But before the upgrade? That's difficult. You need to be Identifying High Risk SQL Before The Upgrade and your company will benefit.

Comparing SQL elapsed times is simply not good enough. Any differences between production and test makes this approach futile. Why? Because there are different business scenarios, workloads, Oracle releases, application releases, hardware configurations... you get the idea.

In this seminar Craig aggressively deals with these complex issues, regardless of the upgrade type. If you are involved with helping to ensure a smooth upgrade, you need to put into practice what Craig teaches in this seminar. It's practical, insightful and a lot of fun.

Members Login Become A Member Today

What You Will Learn

  • Why SQL elapsed times can be different on "similar" systems
  • How to use workload intensity and database time to ID at-risk SQL
  • Why "similar" systems can have very different workload intensities and database time profiles
  • How to quickly profile two systems using a time-based methodology
  • How to properly compare SQL elapsed times on systems with different CPUs
  • How to gather CPU "speed" data
  • How to standardize SQL elapsed times for appropriate comparison
  • How workload intensity factors into SQL elapsed times
  • How different system workloads can falsely tag SQL as risky
  • How different system workloads can miss risky SQL
  • Why functionally focused benchmarks make SQL elapsed time comparisons unreliable
  • How to identify risky situations when the same SQL is used in different business scenarios, runs with different bind variables and uses different interim table names
  • How to deal with situations when a benchmark runs overlapping business scenarios at the same time
  • How to develop a statistical "fingerprint" based on module, action or something else
  • How to pull elapsed time data from AWR tables
  • How to use the free statistical package R to compare two (think: PROD and QAT) elapsed time fingerprints
  • How to use the statistical "fingerprint" results to identify a high risk situation


  • The Truth Hurts: Production vs QAT Realities
  • How To "Fingerprint" An Oracle System For Comparison
  • How To Compare SQL On Different CPU Speed Systems
  • How To Compare SQL On Different Workload Systems
  • How To ID Risk In Highly Business Focused Benchmarks - 1
  • How To ID Risk In Highly Business Focused Benchmarks - 2

Seminar Videos