- Very important fact:
- The size (= # tuples) of the
result set of
a Relational Operator
is not dependent
on the
implementation algorithm !!
Example:
I.e.:
- The output of
T ⋈ U is
the same using
any join algorithm,
such as:
- One-pass join
- Two-pass join based on TPMMS
- Two-pass join based on hashing
- and so on !!!
because the different algorithms implements the same operation !!!
-
The differences between the
algorithms are:
- running time
- memory requirment
Not the output !!!
- The output of
T ⋈ U is
the same using
any join algorithm,
such as:
- The size (= # tuples) of the
result set of
a Relational Operator
is not dependent
on the
implementation algorithm !!
- Note:
- The size of the
result in the
intermediate outputs
do
depend on:
- The order of the operations in the logical query plan
We will see this soon.
- The size of the
result in the
intermediate outputs
do
depend on: