Package com.carrotsearch.hppc

Class ByteHashSet

    • Field Summary

      Fields 
      Modifier and Type Field Description
      protected int assigned
      The number of stored keys (assigned key slots), excluding the special "empty" key, if any.
      protected boolean hasEmptyKey
      Special treatment for the "empty slot" key marker.
      protected int keyMixer
      We perturb hash values with a container-unique seed to avoid problems with nearly-sorted-by-hash values on iterations.
      byte[] keys
      The hash array holding keys.
      protected double loadFactor
      The load factor for keys.
      protected int mask
      Mask for slot scans in keys.
      protected HashOrderMixingStrategy orderMixer
      Per-instance hash order mixing strategy.
      protected int resizeAt
      Expand (rehash) keys when assigned hits this value.

    • Method Summary

      All Methods Static Methods Instance Methods Concrete Methods 
      Modifier and Type Method Description
      boolean add​(byte key)
      Adds k to the set.
      int addAll​(byte... elements)
      Adds all elements from the given list (vararg) to this set.
      int addAll​(ByteContainer container)
      Adds all elements from the given ByteContainer to this set.
      int addAll​(java.lang.Iterable<? extends ByteCursor> iterable)
      Adds all elements from the given iterable to this set.
      protected void allocateBuffers​(int arraySize)
      Allocate new internal buffers.
      protected void allocateThenInsertThenRehash​(int slot, byte pendingKey)
      This method is invoked when there is a new key to be inserted into the buffer but there is not enough empty slots to do so.
      void clear()
      Removes all elements from this collection.
      ByteHashSet clone()
      boolean contains​(byte key)
      Lookup a given element in the container.
      void ensureCapacity​(int expectedElements)
      Ensure this container can hold at least the given number of elements without resizing its buffers.
      boolean equals​(java.lang.Object obj)
      <T extends ByteProcedure>
      T      		 forEach​(T procedure)
      Applies a procedure to all container elements.
      static ByteHashSet from​(byte... elements)
      Create a set from a variable number of arguments or an array of byte.
      int hashCode()
      protected int hashKey​(byte key)
      Returns a hash code for the given key.
      boolean indexExists​(int index)  
      byte indexGet​(int index)
      Returns the exact value of the existing key.
      void indexInsert​(int index, byte key)
      Inserts a key for an index that is not present in the set.
      int indexOf​(byte key)
      Returns a logical "index" of a given key that can be used to speed up follow-up logic in certain scenarios (conditional logic).
      byte indexReplace​(int index, byte equivalentKey)
      Replaces the existing equivalent key with the given one and returns any previous value stored for that key.
      boolean isEmpty()
      Shortcut for size() == 0.
      java.util.Iterator<ByteCursor> iterator()
      Returns an iterator to a cursor traversing the collection.
      protected void rehash​(byte[] fromKeys)
      Rehash from old buffers to new buffers.
      void release()
      Removes all elements from the collection and additionally releases any internal buffers.
      boolean remove​(byte key)
      An alias for the (preferred) removeAll(byte).
      int removeAll​(byte key)
      Removes all occurrences of e from this collection.
      int removeAll​(BytePredicate predicate)
      Removes all elements in this collection for which the given predicate returns true.
      private boolean sameKeys​(ByteSet other)
      Return true if all keys of some other container exist in this container.
      protected void shiftConflictingKeys​(int gapSlot)
      Shift all the slot-conflicting keys allocated to (and including) slot.
      int size()
      Return the current number of elements in this container.
      byte[] toArray()
      Default implementation of copying to an array.
      protected double verifyLoadFactor​(double loadFactor)
      Validate load factor range and return it.
      java.lang.String visualizeKeyDistribution​(int characters)
      Visually depict the distribution of keys.

      • Methods inherited from class java.lang.Object

        finalize, getClass, notify, notifyAll, wait, wait, wait

      • Methods inherited from interface java.lang.Iterable

        forEach, spliterator

    • Field Detail

      • keys

        public byte[] keys
        The hash array holding keys.

      • assigned

        protected int assigned
        The number of stored keys (assigned key slots), excluding the special "empty" key, if any.
        See Also:
        size(), hasEmptyKey

      • mask

        protected int mask
        Mask for slot scans in keys.

      • keyMixer

        protected int keyMixer
        We perturb hash values with a container-unique seed to avoid problems with nearly-sorted-by-hash values on iterations.
        See Also:
        hashKey(byte), "http://issues.carrot2.org/browse/HPPC-80", "http://issues.carrot2.org/browse/HPPC-103"

      • resizeAt

        protected int resizeAt
        Expand (rehash) keys when assigned hits this value.

      • hasEmptyKey

        protected boolean hasEmptyKey
        Special treatment for the "empty slot" key marker.

      • loadFactor

        protected double loadFactor
        The load factor for keys.

    • Method Detail

      • add

        public boolean add​(byte key)
        Adds k to the set.
        Specified by:
        add in interface ByteSet
        Returns:
        Returns true if this element was not part of the set before. Returns false if an equal element is part of the set, and replaces the existing equal element with the argument (if keys are object types).

      • addAll

        public final int addAll​(byte... elements)
        Adds all elements from the given list (vararg) to this set.
        Returns:
        Returns the number of elements actually added as a result of this call (not previously present in the set).

      • addAll

        public int addAll​(ByteContainer container)
        Adds all elements from the given ByteContainer to this set.
        Returns:
        Returns the number of elements actually added as a result of this call (not previously present in the set).

      • addAll

        public int addAll​(java.lang.Iterable<? extends ByteCursor> iterable)
        Adds all elements from the given iterable to this set.
        Returns:
        Returns the number of elements actually added as a result of this call (not previously present in the set).

      • remove

        public boolean remove​(byte key)
        An alias for the (preferred) removeAll(byte).

      • removeAll

        public int removeAll​(byte key)
        Removes all occurrences of e from this collection.
        Specified by:
        removeAll in interface ByteCollection
        Parameters:
        key - Element to be removed from this collection, if present.
        Returns:
        The number of removed elements as a result of this call.

      • removeAll

        public int removeAll​(BytePredicate predicate)
        Removes all elements in this collection for which the given predicate returns true.
        Specified by:
        removeAll in interface ByteCollection
        Returns:
        Returns the number of removed elements.

      • contains

        public boolean contains​(byte key)
        Lookup a given element in the container. This operation has no speed guarantees (may be linear with respect to the size of this container).
        Specified by:
        contains in interface ByteContainer
        Specified by:
        contains in interface ByteLookupContainer
        Returns:
        Returns true if this container has an element equal to e.

      • release

        public void release()
        Removes all elements from the collection and additionally releases any internal buffers. Typically, if the object is to be reused, a simple ByteCollection.clear() should be a better alternative since it'll avoid reallocation.
        Specified by:
        release in interface ByteCollection
        See Also:
        ByteCollection.clear()

      • isEmpty

        public boolean isEmpty()
        Shortcut for size() == 0.
        Specified by:
        isEmpty in interface ByteContainer

      • ensureCapacity

        public void ensureCapacity​(int expectedElements)
        Ensure this container can hold at least the given number of elements without resizing its buffers.
        Specified by:
        ensureCapacity in interface Preallocable
        Parameters:
        expectedElements - The total number of elements, inclusive.

      • size

        public int size()
        Return the current number of elements in this container. The time for calculating the container's size may take O(n) time, although implementing classes should try to maintain the current size and return in constant time.
        Specified by:
        size in interface ByteContainer

      • hashCode

        public int hashCode()
        Overrides:
        hashCode in class java.lang.Object

      • equals

        public boolean equals​(java.lang.Object obj)
        Overrides:
        equals in class java.lang.Object

      • sameKeys

        private boolean sameKeys​(ByteSet other)
        Return true if all keys of some other container exist in this container.

      • clone

        public ByteHashSet clone()
        Overrides:
        clone in class java.lang.Object

      • iterator

        public java.util.Iterator<ByteCursor> iterator()
        Returns an iterator to a cursor traversing the collection. The order of traversal is not defined. More than one cursor may be active at a time. The behavior of iterators is undefined if structural changes are made to the underlying collection.

        The iterator is implemented as a cursor and it returns the same cursor instance on every call to Iterator.next() (to avoid boxing of primitive types). To read the current list's value (or index in the list) use the cursor's public fields. An example is shown below. 

         for (ByteCursor<byte> c : container) {
   System.out.println("index=" + c.index + " value=" + c.value);
 }
        Specified by:
        iterator in interface ByteContainer
        Specified by:
        iterator in interface java.lang.Iterable<ByteCursor>

      • forEach

        public <T extends ByteProcedure> T forEach​(T procedure)
        Applies a procedure to all container elements. Returns the argument (any subclass of ByteProcedure. This lets the caller to call methods of the argument by chaining the call (even if the argument is an anonymous type) to retrieve computed values, for example (IntContainer): 
         int count = container.forEach(new IntProcedure() {
   int count; // this is a field declaration in an anonymous class.
 
   public void apply(int value) {
     count++;
   }
 }).count;
        Specified by:
        forEach in interface ByteContainer

      • forEach

        public <T extends BytePredicate> T forEach​(T predicate)
        Applies a predicate to container elements as long, as the predicate returns true. The iteration is interrupted otherwise.
        Specified by:
        forEach in interface ByteContainer

      • from

        public static ByteHashSet from​(byte... elements)
        Create a set from a variable number of arguments or an array of byte. The elements are copied from the argument to the internal buffer.

      • hashKey

        protected int hashKey​(byte key)
        Returns a hash code for the given key. The default implementation mixes the hash of the key with keyMixer to differentiate hash order of keys between hash containers. Helps alleviate problems resulting from linear conflict resolution in open addressing. The output from this function should evenly distribute keys across the entire integer range.

      • indexOf

        public int indexOf​(byte key)
        Returns a logical "index" of a given key that can be used to speed up follow-up logic in certain scenarios (conditional logic). The semantics of "indexes" are not strictly defined. Indexes may (and typically won't be) contiguous. The index is valid only between modifications (it will not be affected by read-only operations).
        Parameters:
        key - The key to locate in the set.
        Returns:
        A non-negative value of the logical "index" of the key in the set or a negative value if the key did not exist.
        See Also:
        indexExists(int), indexGet(int), indexInsert(int, byte), indexReplace(int, byte)

      • indexExists

        public boolean indexExists​(int index)
        Parameters:
        index - The index of a given key, as returned from indexOf(byte).
        Returns:
        Returns true if the index corresponds to an existing key or false otherwise. This is equivalent to checking whether the index is a positive value (existing keys) or a negative value (non-existing keys).
        See Also:
        indexOf(byte)

      • indexGet

        public byte indexGet​(int index)
        Returns the exact value of the existing key. This method makes sense for sets of objects which define custom key-equality relationship.
        Parameters:
        index - The index of an existing key.
        Returns:
        Returns the equivalent key currently stored in the set.
        Throws:
        java.lang.AssertionError - If assertions are enabled and the index does not correspond to an existing key.
        See Also:
        indexOf(byte)

      • indexReplace

        public byte indexReplace​(int index,
                         byte equivalentKey)
        Replaces the existing equivalent key with the given one and returns any previous value stored for that key.
        Parameters:
        index - The index of an existing key.
        equivalentKey - The key to put in the set as a replacement. Must be equivalent to the key currently stored at the provided index.
        Returns:
        Returns the previous key stored in the set.
        Throws:
        java.lang.AssertionError - If assertions are enabled and the index does not correspond to an existing key.
        See Also:
        indexOf(byte)

      • indexInsert

        public void indexInsert​(int index,
                        byte key)
        Inserts a key for an index that is not present in the set. This method may help in avoiding double recalculation of the key's hash.
        Parameters:
        index - The index of a previously non-existing key, as returned from indexOf(byte).
        Throws:
        java.lang.AssertionError - If assertions are enabled and the index does not correspond to an existing key.
        See Also:
        indexOf(byte)

      • visualizeKeyDistribution

        public java.lang.String visualizeKeyDistribution​(int characters)
        Description copied from interface: ByteSet
        Visually depict the distribution of keys.
        Specified by:
        visualizeKeyDistribution in interface ByteSet
        Parameters:
        characters - The number of characters to "squeeze" the entire buffer into.
        Returns:
        Returns a sequence of characters where '.' depicts an empty fragment of the internal buffer and 'X' depicts full or nearly full capacity within the buffer's range and anything between 1 and 9 is between.

      • verifyLoadFactor

        protected double verifyLoadFactor​(double loadFactor)
        Validate load factor range and return it. Override and suppress if you need insane load factors.

      • rehash

        protected void rehash​(byte[] fromKeys)
        Rehash from old buffers to new buffers.

      • allocateBuffers

        protected void allocateBuffers​(int arraySize)
        Allocate new internal buffers. This method attempts to allocate and assign internal buffers atomically (either allocations succeed or not).

      • allocateThenInsertThenRehash

        protected void allocateThenInsertThenRehash​(int slot,
                                            byte pendingKey)
        This method is invoked when there is a new key to be inserted into the buffer but there is not enough empty slots to do so. New buffers are allocated. If this succeeds, we know we can proceed with rehashing so we assign the pending element to the previous buffer (possibly violating the invariant of having at least one empty slot) and rehash all keys, substituting new buffers at the end.

      • shiftConflictingKeys

        protected void shiftConflictingKeys​(int gapSlot)
        Shift all the slot-conflicting keys allocated to (and including) slot.