pandas set partial data and get SettingWithCopyWarning

Question

I tried to set the partial data to -1, but I get a SettingWithCopyWarning.

I tried to find StackOverflow, but lots of answers use loc to solved.

The data comes from Kaggle Titanic.

import pandas as pd
train = pd.read_csv('data/train.csv')
y = train[["Survived"]]
y.loc[y["Survived"]  == 0,"Survived"] = -1

score 1 · Accepted Answer · answered Mar 26 '18 at 16:55

1

Your logic appears confused. Try this instead:

train.loc[train["Survived"]  == 0,"Survived"] = -1

There is no need to set y = train[['Survived']] and this is what is causing your warning.

You can read about how to use .loc accessor in the Pandas documentation.

answered Mar 26 '18 at 16:55

jpp

But I still don't know why y = train[['Survived']] cause this bug. It is a DataFrame and isn't a copy. Why can't I use loc on y ? – sappy Mar 26 '18 at 17:16
This is a *warning*. It may or may not work, but it's not good practice. – jpp Mar 26 '18 at 17:21
Thanks for your help. I think I know why the warning happened. – sappy Mar 27 '18 at 12:59

1 Answers1