Red-Black tree-based implementation of Map. This class guarantees
that the map will be in both ascending key order and ascending
value order, sorted according to the natural order for the key's
and value's classes.
This Map is intended for applications that need to be able to look
up a key-value pairing by either key or value, and need to do so
with equal efficiency.
While that goal could be accomplished by taking a pair of TreeMaps
and redirecting requests to the appropriate TreeMap (e.g.,
containsKey would be directed to the TreeMap that maps values to
keys, containsValue would be directed to the TreeMap that maps keys
to values), there are problems with that implementation,
particularly when trying to keep the two TreeMaps synchronized with
each other. And if the data contained in the TreeMaps is large, the
cost of redundant storage becomes significant. (See also the new
This solution keeps the data properly synchronized and minimizes
the data storage. The red-black algorithm is based on TreeMap's,
but has been modified to simultaneously map a tree node by key and
by value. This doubles the cost of put operations (but so does
using two TreeMaps), and nearly doubles the cost of remove
operations (there is a savings in that the lookup of the node to be
removed only has to be performed once). And since only one node
contains the key and value, storage is significantly less than that
required by two TreeMaps.
There are some limitations placed on data kept in this Map. The
biggest one is this:
When performing a put operation, neither the key nor the value may
already exist in the Map. In the java.util Map implementations
(HashMap, TreeMap), you can perform a put with an already mapped
key, and neither cares about duplicate values at all ... but this
implementation's put method with throw an IllegalArgumentException
if either the key or the value is already in the Map.
Obviously, that same restriction (and consequence of failing to
heed that restriction) applies to the putAll method.
The Map.Entry instances returned by the appropriate methods will
not allow setValue() and will throw an
UnsupportedOperationException on attempts to call that method.
New methods are added to take advantage of the fact that values are
kept sorted independently of their keys:
Object getKeyForValue(Object value) is the opposite of get; it
takes a value and returns its key, if any.
Object removeValue(Object value) finds and removes the specified
value and returns the now un-used key.
Set entrySetByValue() returns the Map.Entry's in a Set whose
iterator will iterate over the Map.Entry's in ascending order by
their corresponding values.
Set keySetByValue() returns the keys in a Set whose iterator will
iterate over the keys in ascending order by their corresponding
Collection valuesByValue() returns the values in a Collection whose
iterator will iterate over the values in ascending order.