Ankush Patil

Posted on Aug 29

🚀 The Secret Life of Python's range: Lazy, Efficient (Reimplementation Inside)

#computerscience #programming #python

If you’ve been around Python long enough, you’ve probably written something like this:

r = range(10, 100)
print(r)

And then Python hits you with:

range(10, 100)

No list, no numbers, just… vibes. But don’t be fooled — there’s some real wizardry happening here. Let’s peel back the curtain and see how range works under the hood.

🛋️ Lazy by Design

The range object in Python is a lazy sequence. That means:

It doesn’t generate all values upfront.
It computes values on demand.
It sips memory politely instead of guzzling it like a list.

For example:

r = range(1, 10**12)
print(r[0])   # 1
print(r[-1])  # 999999999999

Boom. You just represented a trillion numbers without setting your computer on fire. 🔥

📦 What Does `range` Actually Store?

When you write:

r = range(10, 100, 2)

Python isn’t secretly hiding [10, 12, 14, ..., 98] in your RAM. Instead, it only stores:

start = 10
stop = 100
step = 2
length = 45

That’s it. Four numbers. That’s like fitting an elephant into a matchbox.

🧮 How Indexing Works

Want the 5th element?

print(r[5])

Python does the math:

value = start + (index * step)
value = 10 + (5 * 2) = 20

No loops, no wasted time, just pure arithmetic magic.

Output:

📏 The Length Trick

range can even tell you its length instantly:

print(len(r))

How? Formula time:

$$
length = max(0, \lceil (stop - start) / step \rceil)
$$

Example:

len(range(10, 100, 2))  # 45

No iteration needed. It just knows. 🧙

🔄 Iteration, The Lazy Way

When you do:

for x in range(10, 20, 3):
    print(x)

Python starts at 10, then keeps adding the step until it hits (but doesn’t pass) the stop:

Output:

Just enough work to keep you happy, nothing more.

🛠️ DIY: A Pure Python `range`

Let’s re-create a baby range to see the trick in action:

class MyRange:
    def __init__(self, start, stop=None, step=1):
        if stop is None:
            start, stop = 0, start
        self.start = start
        self.stop = stop
        self.step = step
        self.length = max(0, (stop - start + (step - 1)) // step)

    def __len__(self):
        return self.length

    def __getitem__(self, index):
        if index < 0:
            index += self.length
        if index < 0 or index >= self.length:
            raise IndexError("MyRange index out of range")
        return self.start + index * self.step

    def __iter__(self):
        value = self.start
        for _ in range(self.length):
            yield value
            value += self.step

Test it:

r = MyRange(10, 20, 3)
print(len(r))   # 4
print(r[2])     # 16
print(list(r))  # [10, 13, 16, 19]

Output:

4
16
[10, 13, 16, 19]

Yup, we just cloned range (but worse, because Python’s version is C-powered and lightning fast ⚡).

🕰️ A Brief History Lesson

In Python 2, there were two players:
- range() → made a full list (not lazy).
- xrange() → the lazy version.
In Python 3 (2008), xrange was retired, and range became the cool lazy one.

So when you use range in Python 3 today, you’re really using an evolved xrange.

🎤 Mic Drop

So the next time someone asks you what range(10, 100) is, you can say:

"It’s not a list, it’s a lazy, memory-efficient sequence object that only stores start, stop, step, and calculates values on demand. Also, it’s basically xrange in disguise."

And then walk away dramatically. 🕶️

💡 Have you ever run into a Python behavior that surprised you like this? Drop it in the comments — let’s geek out together!

Top comments (3)

Hamm Gladius • Aug 29

Your writing style is crisp, witty, and incredibly approachable—you turn internals into a fun, memorable read. The metaphors and tight examples (elephant in a matchbox!) made everything click instantly.