The fact that triggers are written in PL/SQL creates a lot of bad publicity for this language because, for many IT professionals, database triggers and performance issues are synonymous. Sadly, there is a lot of truth in this statement. Very few features in the Oracle realm are misused as often as are triggers. The situation is so bad that some of the most outspoken Oracle gurus currently have database triggers included in their lists of things that should never have existed!

This presentation does not take such a radical stance. If you can make DBAs, developers, and architects actually communicate while designing software, the chances of using any feature properly are significantly better. There is nothing “evil” about triggers. They just have to be used where they can actually solve problems.

In this presentation, both table triggers and INSTEAD-OF triggers will be examined from a global system optimization point of view that includes, not only the aspect of functional correctness, but also the tradeoffs between multiple goals. For example, depending upon the available hardware, developers can select either CPU-intense or I/O-intense solutions.

Another key point will be the need to keep your knowledge up-to-date. For example, compound triggers should be used much more frequently for both functional reasons (resolving self-mutation) and performance reasons (bulk operations on tables and simulation of STATEMENT-level behavior on views) This presentation will focus on the most common performance problems related to different kinds of DML triggers and the proper ways of resolving them.