Re: how to build a database from scratch
Date: Sun, 03 Dec 2006 13:10:18 GMT
Message-ID: <_0Ach.441$4p2.103_at_trndny07>
"Murdoc" <murdoc_0_at_hotmail.com> wrote in message news:xn0euizaa3oe8mx000_at_news-south.connect.com.au... Neo wrote:
> > Most relational database systems are based upon Bayer trees.
>
> Is this true? If so why doesn't the relational data model mention trees?
It doesn't need to. B-Trees are not a part of the relational data model, but
are almost
always used as the internal implementation with the RDMS itself. Index
storage is the
perfect example. If you have a multiple-component index, e.g. field_01,
field_02, etc.
then each field would be stored in a separate layer of the B-Tree.
The top layer would contain a separate 'node' for each unique value for
field_01 that
exists. The second layer would contain, for every unique value of field_01,
the
corresponding unique values for field_01. That is, there could exist two
entries in the
second layer for the value "ABC", but they would be related to different
values of
field_01.
That's an incorrect description of how B-trees are used to implement indexes. A single component index will have multiple layers. The division into layers is not determined by the complexity of the index keys (which you called "fields"). It is determined by the ratio of the number of key values in the table to the number of keys that can be pointed to in a single node. Received on Sun Dec 03 2006 - 14:10:18 CET