Storage Indexes - Oracle Exadata Recipes: A Problem-Solution Approach - page 548

Statistic Value (MB)
-------------------------------------------------------------------- -----------------------
cell physical IO bytes eligible for predicate offload 2,254.73
cell physical IO bytes saved by storage index 2,060.34
cell physical IO interconnect bytes 14.01
cell physical IO interconnect bytes returned by smart scan 13.99
Next, we will query a wider range of values:
SQL> select count(*) from d14.myobj_ordered
2 where object_name between 'A' and 'M';
COUNT(*)
----------
9057000
Elapsed: 00:00:00.56
SQL> @lst19-01-exastorind-mysess.sql
Statistic Value (MB)
-------------------------------------------------------------------- -----------------------
cell physical IO bytes eligible for predicate offload 2,254.73
cell physical IO bytes saved by storage index 599.85
cell physical IO interconnect bytes 107.75
cell physical IO interconnect bytes returned by smart scan 107.73
As you can see, when querying wide ranges of values, the storage index I/O savings is not as much as when
selecting a narrower range of values, as expected.
Storage Indexes and Null Values
In this section, we will update roughly 10% of the rows in our test table with null values for cases when
OBJECT_NAME= ' SOURCE$' and execute a query selecting the unmodified rows:
SQL> select count(*) from d14.myobj_ordered
2 where object_name = 'SOURCE$';
COUNT(*)
----------
901
Elapsed: 00:00:00.56
SQL> @lst19-01-exastorind-mysess.sql
Statistic Value (MB)
-------------------------------------------------------------------- -----------------------
cell physical IO bytes eligible for predicate offload 2,254.73
cell physical IO bytes saved by storage index 1,007.05
cell physical IO interconnect bytes .22
cell physical IO interconnect bytes returned by smart scan .19