java.util.ConcurrentModificationException

Where ConcurrentModificationException raised

Case 1: update value Entries during traverse Map in same Thread

java.util.ConcurrentModificationException will be raised in following java code:

 

		Map<String, String> pendingCorrelations = new Hashtable<String, String>();
		pendingCorrelations.put("a", "a1");
		pendingCorrelations.put("b", "b1");
		pendingCorrelations.put("c", "c1");
		pendingCorrelations.put("d", "d1");
		pendingCorrelations.put("e", "e1");
		Set<String> keys = pendingCorrelations.keySet();
		for (String key : keys) { // Here raise ConcurrentModificationException after a is removed.
			if (key.equalsIgnoreCase("a")) {
				 pendingCorrelations.remove(key); 
			}
		}

Exception:

 

java.util.ConcurrentModificationException
	at java.util.Hashtable$Enumerator.next(Hashtable.java:1031)

 

Case 2: This exception could happen in case many simultaneouse threads manipulate same value Object.

 

Why ConcurrentModificationException raised?

Here we focous on two points:

1. How Java analyze for() element?
2. What happened after Hashtable.remove(Object)?

We can image JDK will analyze for() as following:

 

	for(String key=Hashtable$Enumerator.next();Hashtable$Enumerator.hasMoreElements();){
		pendingCorrelations.remove(key);
	}

 

An interesting source code to understand well for() elements:

 

		Map<String, String> pendingCorrelationsMap = new HashMap<String, String>();
		pendingCorrelationsMap.put("c", "c1");
		pendingCorrelationsMap.put("b", "b1");
		pendingCorrelationsMap.put("d", "b1");
		pendingCorrelationsMap.put("a", "a1");

		for (String key : pendingCorrelationsMap.keySet()) {
			if (key.equalsIgnoreCase("a")) { // won't cause ConcurrentModificationException
				pendingCorrelationsMap.remove(key);
			}
		}


		for (String key : pendingCorrelationsMap.keySet()) {
			if (key.equalsIgnoreCase("c")) { // will cause ConcurrentModificationException
				pendingCorrelationsMap.remove(key);
			}
		}

 

 

java.util.Hashtable

 

    /**
     * The number of times this Hashtable has been structurally modified
     * Structural modifications are those that change the number of entries in
     * the Hashtable or otherwise modify its internal structure (e.g.,
     * rehash).  This field is used to make iterators on Collection-views of
     * the Hashtable fail-fast.  (See ConcurrentModificationException).
     */
    private transient int modCount = 0;



	/**
	 * The modCount value that the iterator believes that the backing
	 * Hashtable should have.  If this expectation is violated, the iterator
	 * has detected concurrent modification.
	 */
	protected int expectedModCount = modCount;


    /**
     * Removes the key (and its corresponding value) from this
     * hashtable. This method does nothing if the key is not in the hashtable.
     *
     * @param   key   the key that needs to be removed
     * @return  the value to which the key had been mapped in this hashtable,
     *          or <code>null</code> if the key did not have a mapping
     * @throws  NullPointerException  if the key is <code>null</code>
     */
    public synchronized V remove(Object key) {
	Entry tab[] = table;
	int hash = key.hashCode();
	int index = (hash & 0x7FFFFFFF) % tab.length;
	for (Entry<K,V> e = tab[index], prev = null ; e != null ; prev = e, e = e.next) {
	    if ((e.hash == hash) && e.key.equals(key)) {
		modCount++;  // Only modCount updated within Hashtable.remove(Object)
		if (prev != null) {
		    prev.next = e.next;
		} else {
		    tab[index] = e.next;
		}
		count--;
		V oldValue = e.value;
		e.value = null;
		return oldValue;
	    }
	}
	return null;
    }

We can see that if any Object removed by invode Hashtable.remove(Object), only "modCount++" invoked (not expectedModCount). 

 

java.util.Hashtable$Enumerator<T>

Hashtable$Enumberator<T>

 

    /**
     * A hashtable enumerator class.  This class implements both the
     * Enumeration and Iterator interfaces, but individual instances
     * can be created with the Iterator methods disabled.  This is necessary
     * to avoid unintentionally increasing the capabilities granted a user
     * by passing an Enumeration.
     */
    private class Enumerator<T> implements Enumeration<T>, Iterator<T> {
        ...
	public T next() {
	    if (modCount != expectedModCount)
		throw new ConcurrentModificationException();
	    return nextElement();
	}

	public void remove() {
	    if (!iterator)
		throw new UnsupportedOperationException();
	    if (lastReturned == null)
		throw new IllegalStateException("Hashtable Enumerator");
	    if (modCount != expectedModCount)
		throw new ConcurrentModificationException();

	    synchronized(Hashtable.this) {
		Entry[] tab = Hashtable.this.table;
		int index = (lastReturned.hash & 0x7FFFFFFF) % tab.length;

		for (Entry<K,V> e = tab[index], prev = null; e != null;
		     prev = e, e = e.next) {
		    if (e == lastReturned) {
			modCount++; //both modCount and expectedModCount will be updated simultanously
			expectedModCount++;
			if (prev == null)
			    tab[index] = e.next;
			else
			    prev.next = e.next;
			count--;
			lastReturned = null;
			return;
		    }
		}
		throw new ConcurrentModificationException();
	    }
	}
        ...
}

 

 Hashtable$Enumberator.next() will check (modCount != expectedModCount) to fast fail. Here can see how "java.util.ConcurrentModificationException" raised.

 

Solutions

Solution 1: Update(Remove) java collections through Iterator

As we list in prior java source code, we can see both "modCount" and "expectedCount" will be udpated simultanousely in Hashtable$Enumberator which will be return when calling Hashtable.interator(). 

So we can remove elements as following without ConcurrentModificationException:

 

Iterator iter = Collections.iterator();
while(iter.hasNext()) {
        iter.next(); // Iterator.next() must be called before calling remove()
        iter.remove(); // should be called once and only once
}

Note:

 

1. Here "Collections" is only a symbol of set of objects it is not "java.util.Collection".
2. We can only invoke remove through Iterator. (because there is no other udpate operation API.)

 Solution 2: Remove Objects from Collections outof Traverse Collections.

 

		Map<String, String> pendingCorrelations = new Hashtable<String, String>();
		pendingCorrelations.put("a", "a1");
		pendingCorrelations.put("b", "b1");
		pendingCorrelations.put("c", "c1");
		pendingCorrelations.put("d", "d1");
		pendingCorrelations.put("e", "e1");
		Set<String> keys = pendingCorrelations.keySet();
		List<String> tobeRemvoed = new ArrayList<String>();
		for (String key : keys) {
			if (key.equalsIgnoreCase("a")) {
				tobeRemvoed.add(key);
				// pendingCorrelations.remove(key);
			}
		}

		assertEquals(5, pendingCorrelations.size());
		for (String str : tobeRemvoed) {
			pendingCorrelations.remove(str);
		}
		assertEquals(4, pendingCorrelations.size());