I want to clean NaN values at my data frame separately. I used some filters for finding the NaN values. But same filter detected that another column include NaN values. This condition is confused my mind. I tried a lot of methods but these NaN values don't change.

Firstly see my dataframe;

Raw DataFrame

When the applied filters

pop[(pop['Log GDP per capita'].isna())][(pop['Country name'])=='Somalia']

I used a lot of methods, but I can't change anything.

# Method 1
pop[(pop['Log GDP per capita']).isna()]['Log GDP per capita'].fillna(8,inplace=True)
# Method 2
pop['Log GDP per capita'] = pop['Log GDP per capita'].replace(np.nan,8,inplace=True)
# Method 3
pop[(pop['Log GDP per capita'].isna())][(pop['Country name'])=='Somalia']['Log GDP per capita'].replace(np.nan,7.6,inplace=True)
# Method 4
pop[(pop['Log GDP per capita'].isna())][(pop['Country name'])=='Somalia'].mask(pop['Log GDP per capita']=='', 7.946, inplace=True)
# Method 5
pop[(pop['Log GDP per capita'].isna())][(pop['Country name'])=='Somalia'].mask(pop['Log GDP per capita']==pd.np.nan, 7.946, inplace=True)
# Method 6
pop[(pop['Log GDP per capita'].isna())][(pop['Country name'])=='Somalia'].mask(pop['Log GDP per capita']==np.nan, 7.946, inplace=True)
# Method 7
pop.loc([(pop['Country name']=='Somalia')]['Log GDP per capita'])=7.946 

How can I improve that?

Kaggle Notebook

CodePudding user response：

Use .loc:

mask = pop['Log GDP per capita'].isna() & pop['Country name'].eq('Somalia')
pop.loc[mask, 'Log GDP per capita'] = 8

CodePudding user response：

Maybe you will try to replace Nan values per column

Replace NaN values with specific value per column

Or study id you need this Nan Values in your dataframe