Why does changing a DataFrame in one Jupyter cell also change another variable? [duplicate]

Question

This question already has an answer here:

When does Python Create a New Object (1 answer)

Closed 18 days ago.

I am working in Jupyter Notebook with pandas, and I noticed something strange.

In one cell , I did this:

import pandas as pd
df1 = pd.DataFrame({"A":[1,2,3]})
df2 = df1

Then in another cell, I changed df2:

df2.loc[0,"A"] = 100

But when I check df1 , its also updated - even though I never touched it directly!

print(df1)
output
 A
0 100
1 2
2 3

I expected df1 to stay unchanged. why this happening? Do Jupyter cells share variables differently , or is this pandas work with assignments?

• I tried using df2 = df1.copy() - that seems to fix it.
• I expected df1 and df2 to be two independent DataFrame since I created them separately.
• Just want to understand why the change happens and the right way to avoid it

• Julien: Your expectations are wrong df2 = df1 doesn't create a new object. This just assigns a new variable to the same object. That's it. And you already know the fix: create a copy if you need a copy... And this has nothing to do with jupyter cells.

Answer 1

df2 = df1 only copies the reference, not the data.
So both df1 and df2 point to the same DataFrame in memory — changing one changes the other.

To check their memory addresses:

print("df1:\n", df1)
print("df2:\n", df2)
print("id(df1):", id(df1))
print("id(df2):", id(df2))

df1:
 A
0 100
1 2
2 3
df2:
 A
0 100
1 2
2 3
id(df1): 90949904
id(df2): 90949904

They will be the same.

If you want df2 to be a separate copy, use:

df2 = df1.copy()

Then the ids will be different, and modifying df2 won’t affect df1.