
Chapter 5. Temporal data management and analytics in an operational warehouse 233
This situation is brought into stark relief when we consider rows that have the
following items:
Large objects (BLOBs, CLOBs) or large XML columns
Though rarely updated, they are copied to the history table for each
UPDATE/DELETE on the base table row.
A large number of columns (in the hundreds) with a significant majority that
are rarely or never updated
Again, even though they never change, they are copied over and over to the
history table for each row UPDATE/DELETE.
You can split the base table into two tables:
One table to manage the columns that do change ...