iterator Class

Detailed Description

The QHash::iterator class provides an STL-style non-const iterator for QHash and QMultiHash.

QHash features both STL-style iterators and Java-style iterators. The STL-style iterators are more low-level and more cumbersome to use; on the other hand, they are slightly faster and, for developers who already know STL, have the advantage of familiarity.

QHash<Key, T>::iterator allows you to iterate over a QHash (or QMultiHash) and to modify the value (but not the key) associated with a particular key. If you want to iterate over a const QHash, you should use QHash::const_iterator. It is generally good practice to use QHash::const_iterator on a non-const QHash as well, unless you need to change the QHash through the iterator. Const iterators are slightly faster, and can improve code readability.

The default QHash::iterator constructor creates an uninitialized iterator. You must initialize it using a QHash function like QHash::begin(), QHash::end(), or QHash::find() before you can start iterating. Here's a typical loop that prints all the (key, value) pairs stored in a hash:

QHash<QString, int> hash;
hash.insert("January", 1);
hash.insert("February", 2);
...
hash.insert("December", 12);

QHash<QString, int>::iterator i;
for (i = hash.begin(); i != hash.end(); ++i)
    cout << i.key() << ": " << i.value() << endl;

Unlike QMap, which orders its items by key, QHash stores its items in an arbitrary order. The only guarantee is that items that share the same key (because they were inserted using QHash::insertMulti()) will appear consecutively, from the most recently to the least recently inserted value.

Let's see a few examples of things we can do with a QHash::iterator that we cannot do with a QHash::const_iterator. Here's an example that increments every value stored in the QHash by 2:

QHash<QString, int>::iterator i;
for (i = hash.begin(); i != hash.end(); ++i)
    i.value() += 2;

Here's an example that removes all the items whose key is a string that starts with an underscore character:

QHash<QString, int>::iterator i = hash.begin();
while (i != hash.end()) {
    if (i.key().startsWith('_'))
        i = hash.erase(i);
    else
        ++i;
}

The call to QHash::erase() removes the item pointed to by the iterator from the hash, and returns an iterator to the next item. Here's another way of removing an item while iterating:

QHash<QString, int>::iterator i = hash.begin();
while (i != hash.end()) {
    QHash<QString, int>::iterator prev = i;
    ++i;
    if (prev.key().startsWith('_'))
        hash.erase(prev);
}

It might be tempting to write code like this:

// WRONG
while (i != hash.end()) {
    if (i.key().startsWith('_'))
        hash.erase(i);
    ++i;
}

However, this will potentially crash in ++i, because i is a dangling iterator after the call to erase().

Multiple iterators can be used on the same hash. However, be aware that any modification performed directly on the QHash has the potential of dramatically changing the order in which the items are stored in the hash, as they might cause QHash to rehash its internal data structure. There is one notable exception: QHash::erase(). This function can safely be called while iterating, and won't affect the order of items in the hash. If you need to keep iterators over a long period of time, we recommend that you use QMap rather than QHash.

Warning: Iterators on implicitly shared containers do not work exactly like STL-iterators. You should avoid copying a container while iterators are active on that container. For more information, read Implicit sharing iterator problem.