Nested Tables versus Index-by Tables

The Index-by tables and nested tables are just similar. For e.g.  They have similar structure and their individual elements are accessed in the similar way (by using subscript notation). The main distinction is that the nested tables can be stored in a database column (and hence the word "nested table") while the index-by tables cannot.

The Nested tables extend the functionality of the index-by tables by letting you SELECT,
INSERT, DELETE, and UPDATE nested tables stored in the database. (Keep in mind, that index-by tables cannot be stored in the database). Some collection methods also operate only on the nested tables and varrays. For example, the built-in procedure TRIM cannot be applied to the index-by tables.

Another merit of the nested tables is that an uninitialized nested table is automatically null (that is, the table itself is null, not its elements), while an uninitialized index-by table is simply empty. Therefore, you can apply the IS NULL comparison operator to the nested tables but not to index-by tables.

Though, index-by tables also have some merits. For example, the PL/SQL supports implicit (automatic) datatype conversion between the host arrays and index-by tables (but not nested tables). Therefore, the most efficient way to pass collections to and from the database server is to use the anonymous PL/SQL blocks to bulk-bind input and output host arrays to the index-by tables.