Huffman Encoding, Output bits

Last week I had to write a version of the Huffman encoding / Compression Algorithm. I managed to hack it together. It took perhaps 5 pages of printed code. I am wondering is someone has code they would be willing to share that is much more efficient. (NOTE: this is not for a project that I am submitting - rather now for my own edification.) My university has switched to java from c++ and I am having to teach myself a lot since I started in C++; I am having trouble with the various things involved.

I am well versed on the theory behind it and can do it quite well on paper, it is the efficient java form that escapes me. I have found numerous examples of code online, but none that really fit my specific question well enough.

What I am looking for: an efficient method to read a string in from a file, calculate frequency, construct binary tree with values, traverse the tree and create prefix codes. Output (in a new file) the bit values.

What I think I understand: Input and Output is fairly simple as it isn't that much different from other languages I use. I understand the concept of parsing but haven't had any luck implementing (in the program I submitted I created a variable for each expected value and then just incremented it).

Thanks in advance,
NML

(again I understand if people avoid this for suspicion reasons.)