我将使用shift、cumsum和where的布尔运算：

# Identify 1s
m1 = df['code'].eq(1)
# Identify first 1 of each stretch
m2 = m1&~m1.shift(fill_value=False)

# compute the increment, mask non-1s with 0
df['Period'] = m2.cumsum().where(m1, 0)

输出量：

name  code  Period
0  ray     1       1
1  ray     0       0
2  ray     1       2
3  ray     1       2
4  ray     0       0
5  ray     1       3

中间体：

name  code     m1  ~m1.shift     m2  m2.cumsum  Period
0  ray     1   True       True   True          1       1
1  ray     0  False      False  False          1       0
2  ray     1   True       True   True          2       2
3  ray     1   True      False  False          2       2
4  ray     0  False      False  False          2       0
5  ray     1   True       True   True          3       3

修复代码

您需要跟踪最后一个值。

current_period = 0
last_period = 0
for i in range(len(df)): 
    if df.loc[i, 'code'] == 1:
        if last_period == 0:
            current_period += 1
            last_period = 1
        df.loc[i, 'Period'] = current_period
    else: 
        last_period = 0

但是，你不应该对pandas使用循环。