다음을 통해 공유

What is a hobt_id in Sql 2005?

  • 아티클

The same customer that promted the post of mine titled "How can I get sysindexes information in Sql 2005?" posed a follow-up question - what is the hobt_id column in the sys.partitions catalog view and how does it relate to other catalog views/structures?

A very good question...first, hobt (pronounced "hobbit") stands for "Heap or B-Tree".  Sql Server uses Heap structures to store data when a given table has no cluster (i.e. a clustered index), and uses a B-Tree structure (B-Tree stands for "Balanced Tree", not binary tree as is a common misconception) for indexes (including clustered indexes). A Heap is a collection of data pages that have no specific order and are not linked in any manner (like clusters which are a doubly-linked list).

So, a hobt in Sql Server 2005 is basically a collection of data/index pages - specifically, hobt is generally used to describe a subset of data/index pages within a single partition. Currently, the relationship between a partition and a hobt is ALWAYS 1-to-1 - meaning the hobt_id and partition_id value in the sys.partitions table are completely interchangable at the moment. In fact, you'll notice that the actual values for partition_id and hobt_id are always exactly the same currently in Sql 2005.  In some future release of Sql this may change, however there's no talk of it at this point to my knowledge anyhow.

Chad Boyd ~~~ This posting is provided "AS IS" with no warranties, and confers no rights. Use of any included script samples are subject to the terms specified at https://www.microsoft.com/info/cpyright.htm.

Comments