Updating a value in a Pandas dataframe seems to update all dataframes

Question

I've built two Pandas dataframes like this:

import panda as pd
d = {'FIPS' : pd.Series(['01001', '01002']), 'count' : pd.Series([3, 4])}
df1  = pd.DataFrame(d)
df2 = df1

I want to change one of the values in df2. This is what I've tried:

df2.loc[df2['FIPS'] == '01001','FIPS'] = '01003'

This line appears to update both df1 and df2, but I don't understand why.

Reid, if Jan's reply below answered your question, then accept it as the answer. — zelusp
– zelusp, Commented Aug 24, 2017 at 16:43

Jan Sila · Accepted Answer · 2017-02-04 18:58:49Z

1

Because df2 is only a reference of df1. They point to the same object in the memory, only by different name. If you do df2=df1.copy() it should create a new memory for df2 and only update it..plus you have a typo in import pandas :)

You can check what memory address the object is located in with id(df1) and see its same as df2 and changes if you use the .copy() method

Welcome to SO!

answered Feb 4, 2017 at 18:58

Jan Sila

1,6033 gold badges19 silver badges37 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Reid Over a year ago

Thanks so much! This answers my question. I had no idea that assigning a dataframe to a variable name was so different from assigning an integer.

Jan Sila Over a year ago

Great! :) You can accept my answer by clicking the tick button next to it. Enjoy SO community

Alex L · Accepted Answer · 2017-02-04 18:57:49Z

0

Instead of df2 = df1, say df2 = df1.copy().

The issue is that variables in python act like "pointers" when you assign them complex data structures. They store references to their values, rather than the actual values. So in your code above, df2 becomes another name or alias for df1. Hence the unexpected change.

answered Feb 4, 2017 at 18:57

Alex L

1,1148 silver badges11 bronze badges

Collectives™ on Stack Overflow

Updating a value in a Pandas dataframe seems to update all dataframes

2 Answers 2

2 Comments

Comments

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Comments

Related