Dropping nan string-columns in panda dataframe

Question

I would like to drop all "nan" columns (like the first one on the image) as it doesn't contain any information. I tried this using

df.dropna(how='all', axis=1, inplace=True) 

which unfortunately had no effect. I am afraid that this might be the case because I had to convert my df into a string using

df = df.applymap(str)

This Thread suggests that dropna won't work in such a case, which makes sense to me. I tried to loop over the columns using:

for i in range(len(list(df))):
    if df.iloc[:,i].str.contains('nan').all():
        df.drop(columns=i,axis=1, inplace=True)

which doesn't seem to work. Any help how to drop those columns (and rows as that also doesn't work) is much appreciated

Answer 1

IIUC, try:

df.replace('nan', np.nan).dropna(how='all', axis=1, inplace=True)

This will replace those string 'nan' with np.nan allowing dropna to function as expected.