LeetCode-30-串联所有单词的子串

题目

给定一个字符串 s 和一个字符串数组 words words 中所有字符串 长度相同

s 中的 串联子串 是指一个包含 words 中所有字符串以任意顺序排列连接起来的子串。

  • 例如,如果 words = ["ab","cd","ef"], 那么 "abcdef""abefcd""cdabef""cdefab""efabcd", 和 "efcdab" 都是串联子串。 "acdbef" 不是串联子串,因为他不是任何 words 排列的连接。

返回所有串联子串在 s 中的开始索引。你可以以 任意顺序 返回答案。

示例 1:

1
2
3
4
5
6
输入:s = "barfoothefoobarman", words = ["foo","bar"]
输出:[0,9]
解释:因为 words.length == 2 同时 words[i].length == 3,连接的子字符串的长度必须为 6
子串 "barfoo" 开始位置是 0。它是 words 中以 ["bar","foo"] 顺序排列的连接。
子串 "foobar" 开始位置是 9。它是 words 中以 ["foo","bar"] 顺序排列的连接。
输出顺序无关紧要。返回 [9,0] 也是可以的。

示例 2:

1
2
3
4
5
输入:s = "wordgoodgoodgoodbestword", words = ["word","good","best","word"]
输出:[]
解释:因为 words.length == 4 并且 words[i].length == 4,所以串联子串的长度必须为 16
s 中没有子串长度为 16 并且等于 words 的任何顺序排列的连接。
所以我们返回一个空数组。

示例 3:

1
2
3
4
5
6
输入:s = "barfoofoobarthefoobarman", words = ["bar","foo","the"]
输出:[6,9,12]
解释:因为 words.length == 3 并且 words[i].length == 3,所以串联子串的长度必须为 9
子串 "foobarthe" 开始位置是 6。它是 words 中以 ["foo","bar","the"] 顺序排列的连接。
子串 "barthefoo" 开始位置是 9。它是 words 中以 ["bar","the","foo"] 顺序排列的连接。
子串 "thefoobar" 开始位置是 12。它是 words 中以 ["the","foo","bar"] 顺序排列的连接。

提示:

  • 1 <= s.length <= 104
  • 1 <= words.length <= 5000
  • 1 <= words[i].length <= 30
  • words[i]s 由小写英文字母组成

题解

我的思路跟佬的解法其实是一致的,但是没有选择走下去,因为感觉写不出来,orz想直接跳过难题了(我的思路和朴素哈希表的思路是一致的)

30. 串联所有单词的子串 - 力扣(LeetCode)

  • 用一个哈希表map储存每个word及其出现的次数,单词个数为m,单词长度为w
  • 遍历s,当前位置i为起点,子串sub长度为w * m
  • 然后我们统计单词出现次数是否是正确的,如果正确就继续往后找直到sub末尾,然后记录到ans中;
  • 如果出现没有出现或者次数超过的单词,可以直接剪枝了

事实上,我们可以优化这个枚举起点的过程。

我们可以将起点根据 当前下标与单词长度的取余结果 进行分类,这样我们就不用频繁的建立新的哈希表和进行单词统计。

对于取余结果相同的位置,我们可以运用「滑动窗口」来进行词频统计,复杂度为O(n),当处理完所有余数情况时,代表我们枚举完了所有的起点。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
class Solution {
public List<Integer> findSubstring(String s, String[] words) {
int n = s.length(), m = words.length, w = words[0].length();
// 统计 words 中「每个目标单词」的出现次数
Map<String, Integer> map = new HashMap<>();
for (String str : words) map.put(str, map.getOrDefault(str, 0) + 1);
List<Integer> ans = new ArrayList<>();
for (int i = 0; i < w; i++) {
// 构建一个当前子串对应的哈希表,统计当前子串中「每个目标单词」的出现次数
Map<String, Integer> temp = new HashMap<>();
// 滑动窗口的大小固定是 m * w,每次将下一个单词添加进 temp,上一个单词移出 temp
for (int j = i; j + w <= n; j += w) {
String cur = s.substring(j, j + w);
temp.put(cur, temp.getOrDefault(cur, 0) + 1);
if (j >= i + (m * w)) {
int idx = j - m * w;
String prev = s.substring(idx, idx + w);
if (temp.get(prev) == 1) temp.remove(prev);
else temp.put(prev, temp.get(prev) - 1);
if (!temp.getOrDefault(prev, 0).equals(map.getOrDefault(prev, 0))) continue;
}
if (!temp.getOrDefault(cur, 0).equals(map.getOrDefault(cur, 0))) continue;
// 上面两个 continue 可以减少 map 之间的 equals 操作
if (temp.equals(map)) ans.add(j - (m - 1) * w);
}
}
return ans;
}
}

LeetCode-30-串联所有单词的子串
https://excelius.xyz/leetcode-30-串联所有单词的子串/
作者
Ther
发布于
2024年6月24日
许可协议