Re: Data Dictionary: relative_fno question - Thank You very much !

Thank You, Xuequn and Richard, that enlightened me =) ... Now I know where to go further - great !

Jan =)

"Richard Foote" <richard.foote_at_bigpond.com> schrieb im Newsbeitrag news:69f6c1c8.0301152101.5a0b5c1a_at_posting.google.com...
> xux_at_informa.bio.caltech.edu (Xuequn Xu) wrote in message
news:<b04lk3$q0d$1_at_naig.caltech.edu>...
> > This is normal. The relative_fno (as well as the header_file column) in
> > dba_segments view indicates the file number of the datafile that the
> > segment header resides, and as you know one segment can span more than
one
> > data file. Those "missing" FNO in the view simply means those datafiles
> > have no segment header - they are occupied by the "body" of data
segments
> > whose header is elsewhere (in other FNOs). If you really want to get a
full
> > list of FNOs (for datafiles that has data in them), you can check the
> > dba_extents view - go one step further in a more detail level. A
> > select distinct relative_fno from dba_extents will get you a full list
> > of FNOs as you see from dba_data_files (as long as every datafile has
data,
> > i.e. extents, in it).
> >
>
> Hi Jan,
>
> Just to add to what Xuequn has said, note that Oracle places the first
> extent (which obviously contains the segment header) into the first
> data file listed in the tablespace (based on rfn). This means that
> unless the first data file is full (or sufficiently full not to be
> able to cope with the first extent size), all segment headers will be
> found in this first datafile, hence giving you this result.
>
> Scale away !!
>
> Richard
Received on Thu Jan 16 2003 - 11:56:39 CST