我需要按id和flag对以下数据帧（df2）进行分组，
累积r1 = {'id':'1', 'date':'2020-01-20', 'flag':1}
r2 = {'id':'1', 'date':'2020-01-20', 'flag':1}
r3 = {'id':'1', 'date':'2020-01-20', 'flag':1}
r4 = {'id':'1', 'date':'2020-01-23', 'flag':1}
r5 = {'id':'1', 'date':'2020-01-23', 'flag':1}
r6 = {'id':'1', 'date':'2020-01-22', 'flag':1}
r7 = {'id':'2', 'date':'2019-11-18', 'flag':1}
r8 = {'id':'2', 'date':'2019-11-18', 'flag':1}
r9 = {'id':'2', 'date':'2019-11-21', 'flag':1}
r10 = {'id':'2', 'date':'2019-11-19', 'flag':1}
r11 = {'id':'2', 'date':'2019-11-19', 'flag':1}
r12 = {'id':'2', 'date':'2019-11-19', 'flag':1}
r13 = {'id':'2', 'date':'2019-11-19', 'flag':1}
r14 = {'id':'2', 'date':'2019-11-07', 'flag':1}
df2 = pd.DataFrame([r1,r2,r3,r4,r5,r6,r7,r8,r9,r10,r11,r12,r13,r14])
df2 = df2.set_index('date')
df2
id flag
date
2020-01-20 1 1
2020-01-20 1 1
2020-01-20 1 1
2020-01-23 1 1
2020-01-23 1 1
2020-01-22 1 1
2019-11-18 2 1
2019-11-18 2 1
2019-11-21 2 1
2019-11-19 2 1
2019-11-19 2 1
2019-11-19 2 1
2019-11-19 2 1
2019-11-07 2 1
个值作为计数：
flag如何获取此结果（df3），其中date是具有相同的id和r1 = {'id':'1', 'date':'2020-01-20', 'flag':3}
r4 = {'id':'1', 'date':'2020-01-23', 'flag':2}
r6 = {'id':'1', 'date':'2020-01-22', 'flag':1}
r7 = {'id':'2', 'date':'2019-11-18', 'flag':2}
r9 = {'id':'2', 'date':'2019-11-21', 'flag':1}
r10 = {'id':'2', 'date':'2019-11-19', 'flag':4}
r14 = {'id':'2', 'date':'2019-11-07', 'flag':1}
df3 = pd.DataFrame([r1,r4,r6,r7,r9,r10,r14])
df3 = df3.set_index('date')
df3
id flag
date
2020-01-20 1 3
2020-01-23 1 2
2020-01-22 1 1
2019-11-18 2 2
2019-11-21 2 1
2019-11-19 2 4
2019-11-07 2 1
的值的总和：
StringJoiner