Re: a question about Tom's book "Expert one-on-one Oracle" (about Data Block)

From: Jim Kennedy <kennedy-downwithspammersfamily_at_attbi.net>
Date: Sat, 10 Dec 2005 06:15:25 -0800
Message-ID: <JfWdnccCC_cOfwfenZ2dnUVZ_t-dnZ2d@comcast.com>

"Havel Zhang" <havel.zhang_at_gmail.com> wrote in message news:1134223523.732649.164930_at_f14g2000cwb.googlegroups.com...
> hi everyone:
> i have a question about Tom's new book "Expert one-on-one Oracle".
> In Chapter 2, when talk about data files in Oracle system, tom said
> that the data block structure have a "block header", it has a table
> directory which " The table directory, if present, contains information
> about the tables that store rows in this block (data from more than one
> table may be stored on the same block). "
> My question is, the data block belong to a extent which belong to a
> table/index segment, finally segments belong to a tablespace. How can a
> block can contain more than one table's rows?
>
> anyone have answer? can give me some advice?
>
> Thank you.
>
> BR
>
> Havel
>

There are clustered tables. These are tables that are stored together. For example you might have parent and child tables stored together because these are retrieved together. The key for the rows are not repeated. eg emp and pay records which are related by social security number would be: emp record with SSN

   pay 2006 (no SSN)
  pay 2005 (no ssn)
    pay 2004 (no ssn)
em record #2

pay 2006
pay 2005
emp record #3

    pay 2006
    ...
    pay 1998
...

See docs for clustered tables.
Jim Received on Sat Dec 10 2005 - 08:15:25 CST