Re: create table X(...) => insert into obj$ values (... 'X',...)

It's a good quesion.

How long did it take to create the Universe, when time didn't exist until the Universe was created. Something like that anyway.

Anyway: the point is to go to $ORACLE_HOME/rdbms/admin/bsq.sql and have a thorough read. It's this script that creates all the OBJ$, UET$, FET$ segments; it also creates all the v$ views. When you can recite sql.bsq in your sleep you will know (a) you have been doing Oracle too long (b) you are a deeply disturbed person and (c) you have the makings of a DBA.

Regards
HJR "Marcel Kraupp" <marcel.kraupp_at_gmx.ch> wrote in message news:332bb004.0205272053.1e7f177a_at_posting.google.com...
> Hello
>
> I am trying to get into deeper knowledge on the x$, v$ and X$ (such
> as user$, fet$, obj$) tables/views.
>
> Are the following assumptions correct:
>
> x$
> if the sql parser sees an x$ table, it knows, it's not physically
> stored on harddisk (or tablespaces) but rather is referencing
> some bytes in memory. Therefor, it will return the values
> of the memory disguised as table data. They cannot be altered
> (using insert, updated and delete)
>
> v$
> the same goes for v$ views. The parser realizes these are not
> physically stored anywhere but are referencing data that the
> instance stores (or accumulates over time). They somethimes
> can be altered but this will not generate redo nor undo
> (as this is not data on the hd but in the memory)
>
> X$
> These tables are undistinguishable from ordinary (user created) tables.
> They reside on a tablespace (system); accordingly, these tables
> _could_ (but is not allowed to do by oracle inc.) be altered and would
> generate redo/undo if one did.
>
>
> I also understand, that if I issue a statement like
> create table XYZ (a number, b varchar2(34))
> this will somewhere translate into a
> insert into obj$ values (....,'XYZ',...)
> (though I am not 100% sure of this)
>
> But how can the table obj$ be created if this table must be
> around when (not if) the table is created.
>
> Anything?
>
> MK
Received on Tue May 28 2002 - 16:54:57 CDT