Media recovery issue

Question

Hello
Why is media recovery is faster than, normal DML,DDL,etc?
In production database, It took 1 hour to rebuild indexes and move some objects to diffrent tablespace.
However, When I apply the archivelog of these changes to standby, it took about 20 minutes.
What is reason for this?
Cheers

Answer 1

Media Recovery / Physical Standby operations are faster because they actually only write the final results.
For example, in the index rebuild case, the operations doing the SQL are
a. read table, get column values from table
b. sort values
c. if allocated pga/sort_area is insufficient, overflow to disk
d. run final sort pass, if necessary
e. write sorted results to the index segment
(recursive SQLs to allocate extents, update the data dictionary for the index definition also have to be applied, each of these also having to read the existing extent maps and data dictionary before updating them).
However, when you run media recovery
1. Oracle does not have to read the table
2. There is no sort operation
3. Writes to temp are avoided (as Temp operations are nologging and there is nothing in the redo to "replay" against the tempfiles)
4. The index blocks are "updated" directly