Log oracle generates? doubt on rowid

Question

Hi All,
My oracle database : 10g
lets say I fire a commit and after that I insert 100 rows in a table, I have been told that oracle takes all these insert statements data in a log file and once a commit is issued, it goes on and write this data physically on the database file(DBF file), If I want to access this log, could any1 help me how to do that?
PS: I give u a context why I am asking this question since I am an informatica developer and we have 2 options of loading data into the oracle table, first is normal where oracle generates log for all the dml statements untill a commit is fired and the other one is the BULK mode(considered as a performance enhancing option) which actually skips this log that oracle generates(I want more information about this log but since I don't know the name of this log hence this question).
Some1 asked this question in an interview:
Where does oracle store rowid's ? in the database file itself? is a rowid unique for every row in the database regardless of many schemas(every schema on a different tablespace/database file)?
rowid's could it be duplicated in case where a database comprises of only 2 schema's where both these schema's have different tablespaces?
Many thanks
Rahul

Answer 1

Rowid is unique for the rows in a given table. It can be shared by two rows in tables belonging to the same cluster.
Guess you mean the redo log, and your bulk mode is the direct path insert.
So yes, redo log ensures recovery of data, and direct path directly writes on the disk, and does not have to write redo log... but the database should not be in force logging (which is required for data guard replication).
Also note that commit does not write data physically : non commited data can be written on disk, and commited data could not be written to the disk. The synchronization between buffer cache and disk does not depend on transaction mechanism. (Though synchronization is somehow linked to redo log, as redo log switches issues checkpoint)