A dictionary is a collection of key-value pairs. The key is a TQString used for insertion, removal and lookup. The value is a pointer. Dictionaries provide very fast insertion and lookup.
If you want to use non-Unicode, plain 8-bit \fCchar*\fR keys, use the QAsciiDict template. A QDict has the same performance as a QAsciiDict. If you want to have a dictionary that maps TQStrings to TQStrings use TQMap.
The size() of the dictionary is very important. In order to get good performance, you should use a suitably large prime number. Suitable means equal to or larger than the maximum expected number of dictionary items. Size is set in the constructor but may be changed with resize().
.PP
Items are inserted with insert(); 0 pointers cannot be inserted. Items are removed with remove(). All the items in a dictionary can be removed with clear(). The number of items in the dictionary is returned by count(). If the dictionary contains no items isEmpty() returns TRUE. You can change an item's value with replace(). Items are looked up with operator[](), or with find() which return a pointer to the value or 0 if the given key does not exist. You can take an item out of the dictionary with take().
.PP
Calling setAutoDelete(TRUE) for a dictionary tells it to delete items that are removed. The default behaviour is not to delete items when they are removed.
.PP
When an item is inserted, the key is converted (hashed) to an integer index into an internal hash array. This makes lookup very fast.
.PP
Items with equal keys are allowed. When inserting two items with the same key, only the last inserted item will be accessible (last in, first out) until it is removed.
.PP
The QDictIterator class can traverse the dictionary, but only in an arbitrary order. Multiple iterators may independently traverse the same dictionary.
When inserting an item into a dictionary, only the pointer is copied, not the item itself, i.e. a shallow copy is made. It is possible to make the dictionary copy all of the item's data (a deep copy) when an item is inserted. insert() calls the virtual function TQPtrCollection::newItem() for the item to be inserted. Inherit a dictionary and reimplement newItem() if you want deep copies.
When removing a dictionary item, the virtual function TQPtrCollection::deleteItem() is called. QDict's default implementation is to delete the item if auto-deletion is enabled.
<< fields["surname"]->text() << endl; // Prints "Homer Simpson"
.br
.br
fields.remove( "forename" ); // Does not delete the line edit
.br
if ( ! fields["forename"] )
.br
cout << "forename is not in the dictionary" << endl;
.br
.fi
In this example we use a dictionary to keep track of the line edits we're using. We insert each line edit into the dictionary with a unique name and then access the line edits via the dictionary.
In the example we are using the dictionary to provide fast random access to the keys, and we don't care what the values are. The example is used to generate a menu of TQStyles, each with a unique accelerator key (or no accelerator if there are no unused letters left).
We first obtain the list of available styles, then sort them so that the menu items will be ordered alphabetically. Next we create a dictionary of int pointers. The keys in the dictionary are each one character long, representing letters that have been used for accelerators. We iterate through our list of style names. If the first letter of the style name is in the dictionary, i.e. has been used, we iterate over all the characters in the style name to see if we can find a letter that hasn't been used. If we find an unused letter we put the accelerator ampersand (&) in front of it and add that letter to the dictionary. If we can't find an unused letter the style will simply have no accelerator. If the first letter of the style name is not in the dictionary we use it for the accelerator and add it to the dictionary. Finally we create a QAction for each style.
Constructs a dictionary optimized for less than \fIsize\fR entries.
.PP
We recommend setting \fIsize\fR to a suitably large prime number (e.g. a prime that's slightly larger than the expected number of entries). This makes the hash distribution better which will lead to faster lookup.
.PP
If \fIcaseSensitive\fR is TRUE (the default), keys which differ only by case are considered different.
.SH "QDict::QDict ( const QDict<type> & dict )"
Constructs a copy of \fIdict\fR.
.PP
Each item in \fIdict\fR is inserted into this dictionary. Only the pointers are copied (shallow copy).
.SH "QDict::~QDict ()"
Removes all items from the dictionary and destroys it. If setAutoDelete() is TRUE, each value is deleted. All iterators that access this dictionary will be reset.
Assigns \fIdict\fR to this dictionary and returns a reference to this dictionary.
.PP
This dictionary is first cleared, then each item in \fIdict\fR is inserted into this dictionary. Only the pointers are copied (shallow copy), unless newItem() has been reimplemented.
Sets the collection to auto-delete its contents if \fIenable\fR is TRUE and to never delete them if \fIenable\fR is FALSE.
.PP
If auto-deleting is turned on, all the items in a collection are deleted when the collection itself is deleted. This is convenient if the collection has the only pointer to the items.
.PP
The default setting is FALSE, for safety. If you turn it on, be careful about copying the collection - you might find yourself with two collections deleting the same items.
.PP
Note that the auto-delete setting may also affect other functions in subclasses. For example, a subclass that has a remove() function will remove the item from its data structure, and if auto-delete is enabled, will also delete the item.
.PP
See also autoDelete().
.PP
Examples:
.)l grapher/grapher.cpp, scribble/scribble.cpp, and table/bigtable/main.cpp.
.SH "uint QDict::size () const"
Returns the size of the internal hash array (as specified in the constructor).
