0

I wrote this code to replace the nan values by the median of each column according to the value of the variable to explain: 

data_target1 = data[data['target'] == 1]
data_target0 = data[data['target'] == 0]
for col_name in data.columns.values: 
        data_target1.loc[:,col_name] = data_target1[col_name].fillna(data_target1[col_name].median())
        data_target0.loc[:,col_name] = data_target0[col_name].fillna(data_target0[col_name].median())

data = pd.concat([data_vae1,data_vae0], ignore_index=True)
data = data.sort_values(by='IDCLI_CALCULE')
data = data.reset_index(drop=True)
return data

I get during the execution : 

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

I understand the warning, I have read the documentation and yet I can't write a code without the warning appearing... Could someone explain my mistake and give me a solution to solve this problem?

I read these 2 pages and I understood that I had to use.loc to solve my problem but I remain blocked...

1- : Pandas: SettingWithCopyWarning

2- : Pandas SettingWithCopyWarning

It's probably a detail but I can't find the key to solve this little problem.

Thank you in advance

EDIT

After verification I was able to solve my problem using the "copy" method as suggested in the commentary.

Jhon Snow
  • 67
  • 1
  • 6
  • 2
    Do `data[data['target'] == 1].copy()` when defining the `_target` DataFrames – ALollz Mar 22 '19 at 15:49
  • I will try it as soon as possible but can you just explain a bit why the warning is fired ? Cause I don't realy see where is m'y mistake ! Thanks ! – Jhon Snow Mar 22 '19 at 19:51

0 Answers0