Learned yesterday an important lesson on the Oracle OTN XMLDB forum. A question was asked why a count via a full table scan was quicker than the same example while using an XMLIndex. “XMLIndex performance regarding // (any descendant))”. The poster used the // xpath expression to do his search.
As in most cases nowadays it is important to keep your statistics in order. An XMLIndex is not an “index” as such, it is an logical index / domain index, specially designed for use with XMLDB / XML data.
Unlucky enough, I will not be able to attend and I am sure this will be an very interesting mini seminar.
This is not only because Toon is a very gifted and inspiring speaker, but also because the seminar will present a very useful practical methods and theories applicable to us relational practitioners. The content of this seminar, which can also be found on the website of Ivan Pellegrin (www.maestridellarte.com) that organized this event, will have the following sections: