Juriy about your post:
"See the above quote (or better yet, the quoted document). Besides, even if you have the whole table in one single extent, this doesn't neccesserily mean that the data of that table is stored in contigues space on disk! "
I can't absolutly agree with this point, why:
command CREATE TABLESPACE or ALTER TABLESPACE ADD DATAFILE ... - allocate disk space in 1 step.
In this case if we didn't have heavy fargmentation of OS filesystem
before, then probability getting 1 continious set of i-nodes (or small number of continious sets) very high.
Later, if we set auotoincremet for datafile ONLY, fragmentation of
this datafile will increase, it's true.
but, I very often use raw devices for allocate datafiles
(i like it or not its doesn't matter) in most cases i have to do that in rac or ops environments. Fragmentation datafiles on raw devices ususlly 0.
in this situation we may say about segment fragmentation into tablespace.
This is my point of view on a problem of a disk fragmentation
on oracle db.
