I have a dataframe that contains over 200 columns. I need to create two lists that group columns by type. Rather than creating this list manually I tried different methods that did not work:

numeric_transformer = df.dtypes == 'int64'

or:

numeric_transformer = df.apply(lambda x: x if df.dtypes == 'int64' or 'float64' else 
None, axis=1)

or :

categorical_transformer = df.apply(lambda x: x if df[df.dtypes == 'object'] else None, 
axis=1)

Can anyone help me solve this problem?

CodePudding user response：

I believe that for this question you can use df.select_dtypes like this :

numeric_transformer = df.select_dtypes(exclude='object').columns #exclude object types

categorical_transfomer = df.select_dtypes(include='object').columns

Another way to do this would be using the list comprehension :

numeric_transformer = [column for column in df.columns if (df.dtypes[column] == 'float') or (df.dtypes[column] == 'int')]

categorical_transformer = [column for column in df.columns if df.dtypes[column] == 'object']