388. Longest Absolute File Path [Medium]

自己不会做

本题重要的几点：

1. 用stack回溯（按第二段更妙：用数组），且栈顶需是到上一次遍历完的文件夹的累计目录长度

2. '\t'和'\n'这样的是一个字符，不是两个字符，length、index都只占一位

'\t'和'\n'是java中的特殊字符，'\t'代表tab，'\n'代表换行

/**
 * Runtime: 1 ms, faster than 82.93%
 * Memory Usage: 36.7 MB, less than 91.67%
 */
class Solution {
    public int lengthLongestPath(String input) {
        int maxLen = 0;
        LinkedList<Integer> stack = new LinkedList();
        stack.push(0); // make stack not empty in the first place, so that we can use stack.peek() freely
        
        for (String s : input.split("\n")) {
            int currLevel = s.lastIndexOf('\t') + 1; // '\t' is one character
            while (stack.size() - 2 >= currLevel) { // find the upper level
                stack.pop();
            }
            int len = stack.peek() + s.length() - currLevel;
            if (s.contains(".")) { // file
                maxLen = Math.max(maxLen, len + currLevel);
            } else { // directory
                stack.push(len);
            }
        }
        return maxLen;
    } 
}

/**
 * 更快的方法，不用栈，直接用数组记录更高层的累计路径长度，比用栈更灵活
 * Runtime: 0 ms, faster than 100.00%
 * Memory Usage: 37 MB, less than 63.03%
 */
class Solution {
    public int lengthLongestPath(String input) {
        int maxLen = 0;
        String[] files = input.split("\n");
        int[] levels = new int[files.length + 1]; // 最多可能有files.length层，每一层的累积路径长度记录在(level+1)的索引位置处，这样能保证第0层也能拿到前一索引的数据(0)
        
        for (String file : files) {
            int currLevel = file.lastIndexOf('\t') + 1, currLen = levels[currLevel] + file.length() - currLevel;
            if (file.contains(".")) {
                maxLen = Math.max(maxLen, currLen + currLevel);
            } else {
                levels[currLevel + 1] = currLen;
            }
        }
        return maxLen;
    }
}