Comparator and Priority Queues

I'm in the process of coding Huffman Code where I import a file, generate huffman code for each character, then output the binary to a file. To import the characters I am using a scanner that reads each character, puts it in a node that has values of the read character and a frequency of 1. Then, the node is added to a PriorityQueue. Since the Node class has a compareTo method that compares only frequency, how can I implement a comparator to this specific PriorityQueue that compares the characters when sorting in queue?

Literal example: The queue of characters should be sorted as follows:

[A:1][A:1][A:1][B:1][C:1]
Next step:
[A:1][A:2][B:1][C:1]
Final:
[A:3][B:1][C:1]

Here are some snippets:

protected class Node implements Comparable<Node>{
    Character symbol;
    int frequency;
    
    Node left = null;
    Node right = null;
    @Override
    public int compareTo(Node n) {
        return n.frequency < this.frequency ? 1 : (n.frequency == this.frequency ? 0 : -1);
    }
    
    public Node(Character c, int f){
        this.symbol = c;
        this.frequency = f;
    }
    public String toString(){
        return "["+this.symbol +","+this.frequency+"]";
    }

This is the PriorityQueue that needs a custom comparator:

public static PriorityQueue<Node> gatherFrequency(String file) throws Exception{
    File f = new File(file);
    Scanner reader = new Scanner(f);
    PriorityQueue<Node> PQ = new PriorityQueue<Node>();
    while(reader.hasNext()){
        for(int i = 0; i < reader.next().length();i++){
            PQ.add(new Node(reader.next().charAt(0),1));
        }
    }
    if(PQ.size()>1){ //during this loop the nodes should be compared by character value
        while(PQ.size() > 1){
            Node a = PQ.remove();
            Node b = PQ.remove();
            if(a.symbol.compareTo(b.symbol)==0){
                Node c = new Node(a.symbol, a.frequency + b.frequency);
                PQ.add(c);
            }
            else break;
        }
        return PQ;
    }
    return PQ;
    
}

This is the new method I created using a HashMap:

public static Collection<Entry<Character,Integer>> gatherFrequency(String file) throws Exception{
        File f = new File(file);
        Scanner reader = new Scanner(f);
        HashMap<Character, Integer> map = new HashMap<Character, Integer>();
        while(reader.hasNext()){
            for(int i = 0; i < reader.next().length();i++){
                Character key = reader.next().charAt(i);
                if(map.containsKey(reader.next().charAt(i))){
                    int freq = map.get(key);
                    map.put(key, freq+1);
                }
                else{
                    map.put(key, 1);
                }
            }
        }
        return map.entrySet();
    }

    

The standard approach to implementing Huffman trees is to use a hashmap (in Java, you'd probably use a HashMap<Character, Integer>) to count the frequency for each letter, and insert into the priority queue one node for each letter. So when constructing the Huffman tree itself, you start out with a priority queue that is already in the "final" state that you showed. The Huffman algorithm then repeatedly extracts two nodes from the priority queue, constructs a new parent node for those two nodes, and inserts the new node into the priority queue.