How does SQL Server 2008's Hierarchy data type perform compared to using the hierarchy implementation described by Joe Celko here: http://www.intelligententerprise.com/001020/celko.jhtml?
I've used Celko's method in the past with great results - but don't want to implement it for a new project unless it's better than what Microsoft has provided in SQL Server 2008.
So far, I've only found a single reasonably interesting article on the topic. HierarchyId is much faster than any self-made self-referencing table solution:
http://www.sqlservercentral.com/articles/SQL+Server+2008/62204/
I vaguely remember seeing another comparison elsewhere (but can't find it anymore) which came to similar conclusions; using HierarchyId to build up the hierarchy is a bit cumbersome at best, but query performance is stellar.
Marc
来源：https://stackoverflow.com/questions/1045743/sql-server-2008-hierarchy-data-type-performance