Why `df.shape` is a Property but `df.info()` is a Method in Pandas

[Rajeev Bagra]

Context

When working with Pandas DataFrames, beginners often notice something curious:

df.shape
df.info()

Why does one use parentheses and the other doesn’t?
Is this just a style choice, or is there a deeper reason?

Let’s break it down in a clean, conceptual way.

---

Q1. What is the difference between a property and a method in Python?

A property represents something an object has.
A method represents something an object does.

Think of a DataFrame like a book:

Book	DataFrame
Number of pages	df.shape
Printing the table of contents	df.info()

The page count is a fact about the book.
Printing the table of contents is an action.

---

Q2. Why is df.shape a property?

df.shape returns:

(number_of_rows, number_of_columns)

This information is:

• Already stored inside the DataFrame
• Instantly available
• Requires no calculation
• Has no side effects

So Pandas exposes it as a property:

df.shape

You are simply asking:

“What is your shape?”

Not:

“Please compute something.”

---

Q3. Why is df.info() a method?

df.info() does a lot of work:

It:

• Scans every column
• Counts non-null values
• Detects data types
• Estimates memory usage
• Formats and prints a report

This is a process, not a stored value.
It performs an action, so it is written as a method:

df.info()

You are saying:

“Run the info routine.”

---

Q4. Is this just a design choice?

Yes — but a very deliberate one.

Pandas follows this principle:

Use properties for cheap, always-available facts.
Use methods for expensive or procedural operations.

That’s why:

Pandas API	Type	Reason
df.shape	Property	Stored metadata
df.columns	Property	Already known
df.dtypes	Property	Already known
df.info()	Method	Needs computation & printing
df.describe()	Method	Needs to scan data
df.sum()	Method	Needs calculation

---

Q5. Could Pandas have made info a property and shape a method?

Technically — yes.

Python allows this:

@property
def info(self):
    ...

and

def shape(self):
    ...

But that would be bad API design.

Properties should behave like data.
Methods should behave like actions.

Imagine this:

x = df.info

If that suddenly printed output, it would be confusing and dangerous.

---

Final Takeaway

df.shape is a property because it is an inherent, instant fact about the DataFrame.

df.info() is a method because it is an operation that runs code, scans data, and produces output.

This simple design choice makes Pandas feel natural:

• No parentheses → just reading information
• Parentheses → work is being done

That’s why Pandas feels clean, predictable, and Pythonic.

[Author]

Posts