Iterators

• Summary
• Data Type and Member Function Indexes
• Description
• Key to Iterator Requirements
• Requirements for Input Iterators
• Requirements for Output Iterators
• Requirements for Forward Iterators
• Requirements for Bidirectional Iterators
• Requirements for Random Access Iterators

Pointer generalizations for traversal and modification of collections.

None

Iterators are a generalization of pointers that allow a C++ program to uniformly interact with different data structures. The illustration below displays the five iterator categories defined by the standard library, and shows their hierarchical relationship. Because standard library iterator categories are hierarchical, each category includes all the requirements of the categories above it.

Because iterators are used to traverse and access containers, the nature of the container determines what type of iterator it generates. And, because algorithms require specific iterator types as arguments, it is iterators that, for the most part, determine which standard library algorithms can be used with which standard library containers.

To conform to the C++ standard, all container and sequence classes must provide their own iterators. An instance of a container or sequence's iterator may be declared using either of the following:

class name ::iterator
class name ::const_iterator

Containers and sequences must also provide const iterators to the beginning and end of their collections. These may be accessed using the class members, begin() and end().

The semantics of iterators are a generalization of the semantics of C++ pointers. Every template function that takes iterators will work using C++ pointers for processing typed contiguous memory sequences.

Iterators may be constant or mutable depending upon whether the result of the operator* behaves as a reference or as a reference to a constant. Constant iterators cannot satisfy the requirements of an output_iterator.

Every iterator type guarantees that there is an iterator value that points past the last element of a corresponding container. This value is called the past-the-end value. No guarantee is made that this value is dereferencable.

Every function provided by an iterator is required to be realized in amortized constant time.

The following key pertains to the iterator requirements listed below:

The following expressions must be valid for input iterators:

For input iterators, a == b does not imply that ++a == ++b.

Algorithms using input iterators should be single pass algorithms. That is they should not pass through the same iterator twice.

The value of type T does not have to be an lvalue.

The following expressions must be valid for output iterators:

The only valid use for the operator* is on the left hand side of the assignment statement.

Algorithms using output iterators should be single pass algorithms. That is they should not pass through the same iterator twice.

The following expressions must be valid for forward iterators:

Forward iterators have the condition that a == b implies *a== *b.

There are no restrictions on the number of passes an algorithm may make through the structure.

A bidirectional iterator must meet all the requirements for forward iterators. In addition, the following expressions must be valid:

A random access iterator must meet all the requirements for bidirectional iterators. In addition, the following expressions must be valid:

All relational operators return a value convertible to bool.

©Copyright 1996, Rogue Wave Software, Inc.