字符串字符频率计算代码核心计数逻辑的疑问求助

Hey there! Let's break down that confusing line step by step—it's actually a clever trick using character encoding to make frequency counting super efficient, and it's easier to grasp than it looks.

1. How str.charAt(i) - 'a' Maps Characters to Array Indices #

In Java, every char is secretly an integer under the hood, corresponding to its Unicode value (for lowercase English letters, this matches ASCII values perfectly).

• The character 'a' has a Unicode value of 97, 'b' is 98, all the way up to 'z' which is 122. These values are consecutive—no gaps between them.
• When you subtract 'a' from another lowercase letter, you convert it to a 0-25 index:
  • 'a' - 'a' = 0 → maps to freq[0]
  • 'b' - 'a' = 1 → maps to freq[1]
  • ...
  • 'z' - 'a' = 25 → maps to freq[25]

This turns each lowercase letter into a unique index in the freq array—essentially using the array as a tiny, fast hash table (no fancy hash calculations needed!). The SIZE constant here should definitely be set to 26 to cover all 26 lowercase letters.

2. How freq[str.charAt(i) - 'a']++ Tracks Frequencies #

The ++ part is straightforward once you get the index mapping:

• We loop through every character in the string
• Convert the character to its corresponding array index
• Increment the value at that index by 1

Let's use a quick example with the string "abac":

1. First character 'a' → index 0 → freq[0] becomes 1
2. Second character 'b' → index 1 → freq[1] becomes 1
3. Third character 'a' → index 0 → freq[0] jumps to 2
4. Fourth character 'c' → index 2 → freq[2] becomes 1

By the end of the loop, each position in freq holds the number of times its corresponding letter appeared in the string.

3. A Quick Heads Up About Edge Cases #

One important thing to note: this code only works for lowercase English letters. If your string has uppercase letters, numbers, or symbols, str.charAt(i) - 'a' will give you a negative number or a value greater than 25, which will cause an ArrayIndexOutOfBoundsException. For broader character support, you'd want to use a HashMap<Character, Integer> instead.

内容的提问来源于stack exchange，提问作者abhigyan nayak