划分字母区间

给你一个字符串 s 。我们要把这个字符串划分为尽可能多的片段,同一字母最多出现在一个片段中。

注意,划分结果需要满足:将所有划分结果按顺序连接,得到的字符串仍然是 s

返回一个表示每个字符串片段的长度的列表。

示例 1:

1
2
3
4
5
6
输入:s = "ababcbacadefegdehijhklij"
输出:[9,7,8]
解释:
划分结果为 "ababcbaca"、"defegde"、"hijhklij" 。
每个字母最多出现在一个片段中。
像 "ababcbacadefegde", "hijhklij" 这样的划分是错误的,因为划分的片段数较少。

示例 2:

1
2
输入:s = "eccbbbbdec"
输出:[10]

提示:

  • 1 <= s.length <= 500
  • s 仅由小写英文字母组成

题目本质就是合并区间,统计每个字母出现的第一个位置和最后一个位置,将有重叠的区间合并起来。

因为字母的下标是天然有序的,所以只需要统计每个字母出现的最后一个位置就可以。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
class Solution {
public List<Integer> partitionLabels(String s) {
List<Integer> res = new ArrayList<>();
int[] last = new int[26];
for (int i = 0; i < s.length(); i++) {
last[s.charAt(i) - 'a'] = i;
}
int start = 0;
int end = last[s.charAt(0) - 'a'];
for (int i = 0; i < s.length(); i++) {
int l = i;
int r = last[s.charAt(i) - 'a'];
if (end >= l) {
end = Math.max(end, r);
} else {
res.add(end - start + 1);
start = l;
end = r;
}
}
res.add(end - start + 1);
return res;
}
}