An index on a column contains two types of data: the indexed column value and the
ROWID of the row. Each leaf block of a B-tree index contains the indexed data value and the corresponding
ROWID that is used to locate the actual row in the table. The optimizer sorts the index entries by (key,
ROWID). An index scan occurs whenever the database retrieves column values from an index for a specific column or columns. If your SQL statement refers only to the indexed columns, the database retrieves the column values from the index itself and it doesn't need to access the table. If the query refers to columns other than the indexed column(s), the database accesses the table as well in the next step, using the
ROWIDs it reads during the index ...