Given two words (start and end), and a dictionary, find the length of shortest transformation sequence from start to end, such that: Only one letter can be changed at a time Each intermediate word must exist in the dictionary For example, Given: start = "hit" end = "cog" dict = ["hot","dot","dog","lot","log"] As one shortest transformation is "hit" -> "hot" -> "dot" -> "dog" -> "cog", return its length 5. Note: Return 0 if there is no such transformation sequence. All words have the same length. All words contain only lowercase alphabetic characters.
难度:96.这道题看似一个关于字符串操作的题目,其实要解决这个问题得用图的方法。我们先给题目进行图的映射,顶点则是每个字符串,然后两个字符串如果相差一个字符则我们进行连边。接下来看看这个方法的优势,注意到我们的字符集只有小写字母,而且字符串长度固定,假设是L。那么可以注意到每一个字符可以对应的边则有25个(26个小写字母减去自己),那么一个字符串可能存在的边是25*L条。接下来就是检测这些边对应的字符串是否在字典里,就可以得到一个完整的图的结构了。根据题目的要求,等价于求这个图一个顶点到另一个顶点的最短路径,一般我们用广度优先搜索。这里因为需要确定level数作为最短路径跳数,所以类似Binary Tree Level Order Traversal那样记录了上下两层节点数ParentNum 和ChildNum
注意19行字符串转换为String最好用 String st = new String(current); 使用 String st = current.toString(); 要出错
而长度为L的字符串总共有26^L,所以时间复杂度是O(min(26^L, size(dict)),空间上需要存储访问情况,也是O(min(26^L, size(dict))。
1 public class Solution { 2 public int ladderLength(String start, String end, Set<String> dict) { 3 if (start==null || end==null || start.length()==0 || end.length()==0 || start.length()!=end.length()) 4 return 0; 5 HashSet<String> visited = new HashSet<String>(); 6 LinkedList<String> queue = new LinkedList<String>(); 7 queue.add(start); 8 visited.add(start); 9 int parentNum = 1; 10 int childNum = 0; 11 int level = 1; 12 while (!queue.isEmpty()) { 13 String cur = queue.poll(); 14 parentNum--; 15 for (int i=0; i<cur.length(); i++) { 16 char[] current = cur.toCharArray(); 17 for (char c='a'; c<='z'; c++) { 18 current[i] = c; 19 String st = new String(current); 20 if (st.equals(end)) { 21 return level+1; 22 } 23 if (dict.contains(st) && !visited.contains(st)) { 24 queue.add(st); 25 visited.add(st); 26 childNum++; 27 } 28 } 29 } 30 if (parentNum == 0) { 31 parentNum = childNum; 32 childNum = 0; 33 level++; 34 } 35 } 36 return 0; 37 } 38 }
作者道题还是有很多心得体会的,比如String是immutable的,如何改动其中一个元素?上面代码16行提供了一个非常好的方法,那就是先把String转化为CharArray, 改了元素之后再转回String,注意CharArray转String是new String(array). 另一种办法就是用StringBuffer的setCharAt()
写DFS, BFS不要忘了定义一个isVisited数组记录访问过的节点
还曾有另外一种想法:定义一个isWithinOne函数,判断两个String距离是否为1,但是可惜TLE了,主要是因为要挨个遍历dict里面每个元素,dict很大就过不了。不过思路还是很好的
1 public class Solution { 2 public int ladderLength(String start, String end, Set<String> dict) { 3 LinkedList<String> queue = new LinkedList<String>(); 4 HashSet<String> isVisited = new HashSet<String>(); 5 queue.offer(start); 6 isVisited.add(start); 7 int PNum = 1; 8 int CNum = 0; 9 int level = 1; 10 while (!queue.isEmpty()) { 11 String cur = queue.poll(); 12 PNum--; 13 if (isWithinOne(cur, end)) return level+1; 14 for (String item : dict) { 15 if (!isVisited.contains(item) && isWithinOne(cur, item)) { 16 isVisited.add(item); 17 queue.offer(item); 18 CNum++; 19 } 20 } 21 if (PNum == 0) { // end with current level 22 PNum = CNum; 23 CNum = 0; 24 level++; 25 } 26 } 27 return 0; 28 } 29 30 public boolean isWithinOne(String cur, String end) { 31 int diff = 0; 32 for (int i=0; i<cur.length(); i++) { 33 if (cur.charAt(i) != end.charAt(i)) { 34 diff++; 35 } 36 } 37 return diff==1? true : false; 38 } 39 }