如何用JavaScript实现特定规则的数字舍入？

Alright, let's break down your custom rounding problem and fix this properly. First, let's make sure we're on the same page with the rules you need:

• For positive numbers: No matter if the decimal is above or below 0.5, round up to the next closest 0.5 multiple. So 0.523 goes to 1, 2.235 goes to 2.5.
• For negative numbers: Round away from 0 (i.e., down to the next smaller 0.5 multiple). So -0.081 goes to -0.5, -1.081 goes to -1.5.

Why Your calc Method Didn't Work #

Your calc function had two key issues:

1. Wrong granularity: You used 0.00005, which equals 0.005% when working with percentage values. But your rule requires rounding to 0.5% increments—so if you're passing in percentage values (like 0.523 for 0.523%), the granularity should be 0.5. If you're using decimal equivalents (like 0.00523 for 0.523%), it should be 0.005.
2. Incorrect rounding logic: Math.round() uses standard "round half up" behavior, which doesn't match your "always round away from 0" requirement.

A Working Solution (Refined From Your round Function) #

Your original round function was nearly there—let's refine it to be more explicit and handle edge cases (like exact 0.5 multiples that don't need rounding):

static customRound(num) {
  // Return exact 0.5 multiples as-is, no rounding needed
  if (num % 0.5 === 0) {
    return num;
  }

  const absNum = Math.abs(num);
  const sign = num < 0 ? -1 : 1;

  // For positives: round up to the next 0.5 multiple
  // For negatives: this effectively rounds down away from 0
  const roundedAbs = Math.ceil(absNum / 0.5) * 0.5;

  return sign * roundedAbs;
}

Test With Your Examples #

Assuming you're passing in percentage values (e.g., 0.523 represents 0.523%):

• customRound(0.523) → 1 ✅
• customRound(2.235) → 2.5 ✅
• customRound(-0.081) → -0.5 ✅
• customRound(-1.081) → -1.5 ✅

How This Logic Works #

The core idea is simple and flexible:

1. Divide the number by your granularity (0.5) to convert the problem into rounding integers
2. Use Math.ceil() on the absolute value to always round up (for negatives, applying the original sign turns this into rounding away from 0)
3. Multiply back by the granularity to get your final rounded value

If you ever need to adjust the rounding increment (e.g., to 0.25 instead of 0.5), just swap out the 0.5 values with your new desired granularity.

内容的提问来源于stack exchange，提问作者Noah Tony