A value is trying to be set on a copy of a slice from a DataFrame

Question

In 2 different situations I get the warning "A value is trying to be set on a copy of a slice from a DataFrame"

in the code:

# Setup
import os

import pandas as pd

from utils import make_chaos

from tests.helpers import check

pd.options.display.max_rows = 10
users = pd.read_csv(os.path.join('data', 'users.csv'), index_col=0)
# Pay no attention to the person behind the curtain
make_chaos(users, 19, ['first_name'], lambda val: val.lower())



## CHALLENGE - Verified email list ##

# TODO: Narrow list to those that have email verified.
users2 = users[(users.email_verified == True)]

#  The only columns should be first, last and email
email_list = users[:]

users3 = users2[['first_name','last_name','email']]


# TODO: Remove any rows missing last names

users3.dropna(subset = ["last_name"], inplace=True)


users4 = users3.dropna(subset = ["last_name"])



# # TODO: Ensure that the first names are the proper case
users4.loc[users4.first_name.str.islower(), 'first_name'] = users4.first_name.str.title()

# Return the new sorted DataFrame..last name then first name ascending
users4.sort_values(['last_name','first_name'])

I get this error with

users3.dropna(subset = ["last_name"], inplace=True)

and

users4.loc[users4.first_name.str.islower(), 'first_name'] = users4.first_name.str.title()

In the first case I am using inplace which I thought should remedy this issue. In the second case I closely followed the s2n08-manipulating-text.ipynb example. I would greatly appreciate if someone can explain what I am doing wrong.

Answer 1 · 2020-06-25T06:48:26Z

June 25, 2020 6:48am

You are trying to set new values on a view (users3) of users2. Essentially what fixed the issue was to create a copy of the dataframe.

users3 = users2[['first_name','last_name','email']].copy()

This is my code and it took me a while to figure out why yours threw the error. But in my code example I wasn't using the view to change anything. It comes at the end only to be sorted before I display it.

## CHALLENGE - Verified email list ##

# TODO: Narrow list to those that have email verified.
#  The only columns should be first, last and email
email_list = users[:]
email_list = email_list[email_list.email_verified == True]

# TODO: Remove any rows missing last names
email_list.dropna(how='any', inplace=True)

# TODO: Ensure that the first names are the proper case
email_list.loc[email_list.first_name.str.islower(), 'first_name'] = email_list.first_name.str.title()
len(email_list[email_list.first_name.str.islower()])

# Return the new sorted DataFrame..last name then first name ascending
email_list2 = email_list[['first_name', 'last_name', 'email']]
email_list2.sort_values(by=['last_name', 'last_name'])

Answer 2 · 2020-07-13T17:46:00Z

July 13, 2020 5:46pm

## CHALLENGE - Verified email list ##

# TODO: Narrow list to those that have email verified.
verified_emails = users["email_verified"] == True
#  The only columns should be first, last and email
email_list = users.loc[verified_emails, ["first_name","last_name","email_verified"]]

# TODO: Remove any rows missing last names
email_list.dropna(inplace=True)
# TODO: Ensure that the first names are the proper case
email_list.loc[email_list.first_name.str.islower(), "first_name"] = email_list.first_name.str.title()
email_list.loc[email_list.last_name.str.islower(), "last_name"] = email_list.last_name.str.title()

# Return the new sorted DataFrame..last name then first name ascending
email_list.sort_values(["last_name", "first_name", "email_verified"], inplace=True)

#I did email_list.dropna() because the only empty rows were last names```

Welcome to the Treehouse Community

Looking to learn something new?

Sebastiaan van Vugt

Sebastiaan van Vugt

A value is trying to be set on a copy of a slice from a DataFrame

2 Answers

Malte Niepel

Malte Niepel

Mike Trump

Mike Trump