How to use Python Pandas to count multiple column with unique date

Question

I want to count the unique date with multiple column, I use "pd.value_counts" ,but it only can count single column.

import os
import sys
import pandas as pd
import numpy as np
from dateutil.parser import *
from datetime import *

df = pd.read_csv('test.csv',encoding='utf8') 
a= {"date":df.date,
    "good":df.good,
    "bad":df.bad,
    "normal":df.norma
   }
df2=pd.DataFrame(a)
print(df2)

---

output:
         date  bad  normal  good
0  2017-09-16  1    0       0
1  2017-09-16  0    1       0
2  2017-09-20  0    0       1
3  2017-09-20  0    0       1
4  2017-09-22  0    1       0

---

Have any function in Pandas can count unique date like this

         date  bad  normal  good
0  2017-09-16  1    1       0
1  2017-09-20  0    0       2
2  2017-09-22  0    1       0