Python DataFrame obtenga columnas numéricas
dfnew = df.select_dtypes(include=np.number)
Panicky Pony
dfnew = df.select_dtypes(include=np.number)
dfnew = df._get_numeric_data()
cols = df.select_dtypes([np.number]).columns
non_numeric_cols = pd.Index(set(df.columns).difference(set(df.select_dtypes(include=(np.number)).columns)))