Unable to write function for df.columns to factorize() - python-3.x

Question

I have a dataframe df:
age 45211 non-null int64
job 45211 non-null object
marital 45211 non-null object
default 45211 non-null object
balance 45211 non-null int64
housing 45211 non-null object
loan 45211 non-null object
contact 45211 non-null object
day 45211 non-null int64
month 45211 non-null object
duration 45211 non-null int64
campaign 45211 non-null int64
pdays 45211 non-null int64
previous 45211 non-null int64
poutcome 45211 non-null object
conversion 45211 non-null int64
I want to do two things:
(1) I want to create two sub-dataframes which will be automatically separated by dtype=object and dtype=int64. I thought of something like this:
object_df=[]
int_df=[]
for i in df.columns:
if dtype=object:
*add column to object_df*
else:
*add column to int_df*
(2) Next, I want to use the columns from object_df['job','marital','default','housing','loan','contact','month','poutcome'] and write a function which factorizes each column, so that categories will be converted to numbers. I thought of something like this:
job_values,job_labels= df['job'].factorize()
df['job_fac']=job_values
Since I would have to copy and paste those for all columns in the object_df, is there a way to write a neat dynamic function?

Answer 1

Use DataFrame.select_dtypes first:
object_df = df.select_dtypes(object)
int_df = df.select_dtypes(np.number)
And then create lambda function for factorize, DataFrame.add_suffix and DataFrame.join to original DataFrame:
cols = ['job','marital','default','housing','loan','contact','month','poutcome']
df = df.join(object_df[cols].apply(lambda x: pd.factorize(x)[0]).add_suffix('_fac'))
Sample:
np.random.seed(2020)
c = ['age', 'job', 'marital', 'default', 'balance', 'housing', 'loan', 'contact', 'day', 'month', 'duration', 'campaign', 'pdays', 'previous', 'poutcome', 'conversion']
cols = ['job','marital','default','housing','loan','contact','month','poutcome']
cols1 = np.setdiff1d(c, cols)
df1 = pd.DataFrame(np.random.choice(list('abcde'), size=(3, len(cols))), columns=cols)
df2 = pd.DataFrame(np.random.randint(10, size=(3, len(cols1))), columns=cols1)
df = pd.concat([df1, df2], axis=1).reindex(columns=c)
print (df)
age job marital default balance housing loan contact day month duration \
0 9 a a d 5 d d d 6 a 5
1 4 a a c 2 b d d 7 c 1
2 3 a e e 2 a e b 1 b 2
campaign pdays previous poutcome conversion
0 6 4 6 a 6
1 3 4 9 d 4
2 0 7 1 c 9
object_df = df.select_dtypes(object)
print (object_df)
job marital default housing loan contact month poutcome
0 a a d d d d a a
1 a a c b d d c d
2 a e e a e b b c
int_df = df.select_dtypes(np.number)
print (int_df)
age balance day duration campaign pdays previous conversion
0 9 5 6 5 6 4 6 6
1 4 2 7 1 3 4 9 4
2 3 2 1 2 0 7 1 9
cols = ['job','marital','default','housing','loan','contact','month','poutcome']
df = df.join(object_df[cols].apply(lambda x: pd.factorize(x)[0]).add_suffix('_fac'))
print (df)
age job marital default balance housing loan contact day month ... \
0 9 a a d 5 d d d 6 a ...
1 4 a a c 2 b d d 7 c ...
2 3 a e e 2 a e b 1 b ...
poutcome conversion job_fac marital_fac default_fac housing_fac \
0 a 6 0 0 0 0
1 d 4 0 0 1 1
2 c 9 0 1 2 2
loan_fac contact_fac month_fac poutcome_fac
0 0 0 0 0
1 0 0 1 1
2 1 1 2 2