Integers Rescaling: Normalize a List of Integers to a Target Range Without Floating-Point Loss

Question

## Problem

Given a list of integers and a target range [lo, hi], rescale each value to fit within that range. Values should be scaled proportionally. Handle edge cases: all identical values, negative inputs, and integer-only output.

```python
def rescale(values: list[int], lo: int, hi: int) -> list[float]:
    # Maps min(values)->lo, max(values)->hi linearly
    # Returns rescaled values as floats
    pass

def rescale_int(values: list[int], lo: int, hi: int) -> list[int]:
    # Same but rounds to nearest integer
    pass
```

**Example:**
```
rescale([10, 20, 30], 0, 100) -> [0.0, 50.0, 100.0]
rescale([5, 5, 5], 0, 10)    -> [5.0, 5.0, 5.0]  # degenerate case
rescale([-10, 0, 10], 0, 1)  -> [0.0, 0.5, 1.0]
```

## Follow-ups

1. When all values are equal, what value should the output contain? Justify your choice.
2. How does floating-point precision affect the rescaled output when the range is very large?
3. How would you implement a streaming version that updates the scale as new values arrive?
4. How would you extend this to support per-column rescaling on a 2D matrix?

Integers Rescaling: Normalize a List of Integers to a Target Range Without Floating-Point Loss

Question Details

Full Details

Topics

More from Waymo