Which of these takes less time?

Question

Which of these takes less time?
INSERT INTO TABLE_1 AS SELECT * FROM TABLE_2;
CREATE TABLE TABLE_1 SELECT * FROM TABLE_2;
Thanks,
Bhagat

Answer 1

here's what the documentation says about a table created with the NOLOGGING clause:
"Specify NOLOGGING if you do not want these operations to be logged.
For a nonpartitioned object, the value specified for this clause is the actual physical attribute of the segment associated with the object."
I read that to mean that segment information is not logged. I don't see where it says that inserts are not logged.
And my testing on versions 9, 10.1 and 10.2 have shown that normal inserts to a table created with nologging mode will generate the same amount of redo as inserts to a table created with logging mode.
Using APPEND will drastically reduce the amount of redo generated. Here's some of my findings showing the amount of redo (in bytes) generated for different methods:
v10.2.0.3.0
table structure:
t( a number, b number)
insert 50,000 rows per test
created regular table, did NOT enable parallel dml
insert append parallel :  141808 redo
insert parallel        :  142032 redo
insert append          :    1156 redo
insert                 :  754120 redo
regular table, alter session enable parallel dml:
insert append parallel :  141180
insert parallel        :  141876
insert append          :   16156
insert                 :  748552
create table with nologging option, did NOT enable parallel dml:
insert append          :   16000
insert                 :  748616let the debate begin (again)