CODEWITHSUNDEEP
pythonpandasdataframe
Klausos Klausos
Klausos Klausos

Reputation: 16060

Python pandas SettingWithCopyWarning

I wonder why the first code (that reads data from CSV) has a warning message SettingWithCopyWarning: A value is trying to be set on a copy of a slice from a DataFrame, while the second code (that uses the same logic, but manual DataFrame initialization) does not have this warning message.

Code #1

import pandas as pd

myData = pd.read_csv('https://spark-public.s3.amazonaws.com/dataanalysis/loansData.csv')

for i in range(len(myData['Employment.Length'])):
    myData['Employment.Length'][i] = 3

Code #2

list1 = ['bla', 'la', 'lal']    
list2 = [1, 2, 3]      
myData = pd.DataFrame({'1': list1, '2':list2})  

for i in range(len(myData)):  
    myData['1'][i] = 3

Upvotes: 1

Views: 328

Answers (1)

shanmuga
shanmuga

Reputation: 4499

Both the above codes generate the same warning. When both codes are run in same python application/console the warning is generated twice (once from each code) but it is reported only once due to the warnings filter being set to once

once: print only the first occurrence of matching warnings, regardless of location.
always: always print matching warnings

Refer documentation

To change this behavior set warnings filter to always

import warnings
warnings.filterwarnings("always")

Upvotes: 1

Related Questions