如何在python中更改dataframe中的列值？

Question

我用python创建了一个数据帧，

import pandas as pd
d = {'col1': ["day1", "7:00", "8:00","9:00", "10:00", "11:00",
               "day2", "7:00", "8:00","9:00", "10:00", "11:00",
               "day3", "7:00", "8:00","9:00", "10:00", "11:00"],
      'col2': [0, 4.1, 3, 3.5, 45.1, 16.9,
               0, 6.5, 4, 9.8, 33.9, 19.8,
               0, 6.9, 2.5, 7, 81.1, 13.8]}
df = pd.DataFrame(data=d)
print(df)
  col1  col2
0    day1   0.0
1    7:00   4.1
2    8:00   3.0
3    9:00   3.5
4   10:00  45.1
5   11:00  16.9
6    day2   0.0
7    7:00   6.5
8    8:00   4.0
9    9:00   9.8
10  10:00  33.9
11  11:00  19.8
12   day3   0.0
13   7:00   6.9
14   8:00   2.5
15   9:00   7.0
16  10:00  81.1
17  11:00  13.8

例如，我想将col1中的所有这些时间线数据更改为天

 col1  col2
    0    day1   0.0
    1    day1   4.1
    2    day1   3.0
    3    day1   3.5
    4    day1  45.1
    5    day1  16.9
    6    day2   0.0
    7    day2   6.5
    8    day2   4.0
    9    day2   9.8
    10   day2  33.9
    11   day2  19.8
    12   day3   0.0
    13   day3   6.9
    14   day3   2.5
    15   day3   7.0
    16   day3  81.1
    17   day3  13.8

这只是一个样本数据集。因此，我希望有一个简单的通用答案来解决这个问题。例如，如果我们有1000天的数据集..

Answer 1

df.col1=df.col1.where(df.col1.str.isalnum()).ffill()
df
Out[242]: 
    col1  col2
0   day1   0.0
1   day1   4.1
2   day1   3.0
3   day1   3.5
4   day1  45.1
5   day1  16.9
6   day2   0.0
7   day2   6.5
8   day2   4.0
9   day2   9.8
10  day2  33.9
11  day2  19.8
12  day3   0.0
13  day3   6.9
14  day3   2.5
15  day3   7.0
16  day3  81.1
17  day3  13.8

Answer 2

尝试丢弃时间戳并向前填充：

# Remove timestamps 
discard_mask = ~df.col1.str.startswith('day')
df.loc[discard_mask, "col1"] = np.nan

# Forward fill
df.ffill()

#     col1  col2
# 0   day1   0.0
# 1   day1   4.1
# 2   day1   3.0
# 3   day1   3.5
# 4   day1  45.1
# 5   day1  16.9
# 6   day2   0.0
# 7   day2   6.5
# 8   day2   4.0
# 9   day2   9.8
# 10  day2  33.9
# 11  day2  19.8
# 12  day3   0.0
# 13  day3   6.9
# 14  day3   2.5
# 15  day3   7.0
# 16  day3  81.1
# 17  day3  13.8