table() add and remove values - algorithm

Question

I'm trying to create an efficient table() calculation (get the frequency of each value in a vector). The difference from the ordinary table() function is that it needs to support adding and removing values without recalculating the whole table.
I thought of using a hash table. for add: look for the key, add 1 to the value. for remove: look for the key. if found: subtract 1 from value, if not found: add new key with value=1.
I was wondering if any of you have other ideas.
Example:
X
key freq
1 3
2 5
3 2
8 1
remove(8)
key freq
1 3
2 5
3 2
add(2)
key freq
1 3
2 6
3 2
Any ideas for an efficient implementation?
Thanks in advance!
--EDIT--
My current code, if anyone is interested (also involves the calculation of shannon entropy)
create.freq.hash<-function(x)
{
t<-table(x)
h<-hash(names(t),as.numeric(t));
return(h);
}
freq.hash.add<-function(hash,key)
{
if(is.null(hash[[key]]))
{
.set(hash,key,+1)
}
else
{
.set(hash,key,hash[[key]]+1)
}
}
freq.hash.remove<-function(hash,key)
{
if(!is.null(hash[[key]]))
{
if(hash[[key]]==1)
del(key,hash)
else
.set(hash,key,hash[[key]]-1)
}
}
hash.entropy<-function(hash)
{
if(is.empty(hash))
return;
v<-values(hash);
v.prob<-v/sum(v);
entropy = (-1)*(v.prob%*%log2(v.prob))
return(entropy)
}