Question

下面的数据框由pd.read_sql填充。如何为book_date == start_date中的每个唯一Group / SubGroup对选择wf值，并将其存储在“new”列中。

*为了更加清晰，我对行进行了星号，星号不在数据集中。

|       | Group | SubGroup  | book_date | start_date | wf   | co2      | new |
|-------|-------|-----------|-----------|------------|------|----------|-----|
| 236   | Virgo | Milkyway  | 3/1/1985  | 5/1/1985   | 0.04 | NaN      |     |
| 239   | Virgo | Milkyway  | 4/1/1985  | 5/1/1985   | 0.05 | NaN      |     |
| 1178  | Virgo*| Milkyway* | 5/1/1985* | 5/1/1985*  | 0.06*| 0.004179*|     |
| 535   | Virgo | Milkyway  | 6/1/1985  | 5/1/1985   | 0.07 | 0.008245 |     |
| 1056  | Virgo | Andromeda | 6/1/1993  | 8/1/1993   | 1.57 | NaN      |     |
| 1046  | Virgo | Andromeda | 7/1/1993  | 8/1/1993   | 1.58 | NaN      |     |
| 956   | Virgo*| Andromeda*| 8/1/1993* | 8/1/1993*  | 1.59*| 0.006688*|     |
| 776   | Virgo | Andromeda | 9/1/1993  | 8/1/1993   | 1.60 | 0.012917 |     |

这是预期的结果。

|       | Group | SubGroup  | book_date | start_date | wf   | co2      | new  |
|-------|-------|-----------|-----------|------------|------|----------|------|
| 236   | Virgo | Milkyway  | 3/1/1985  | 5/1/1985   | 0.04 | NaN      | 0.06 |
| 239   | Virgo | Milkyway  | 4/1/1985  | 5/1/1985   | 0.05 | NaN      | 0.06 |
| 1178  | Virgo*| Milkyway* | 5/1/1985* | 5/1/1985*  | 0.06*| 0.004179*| 0.06 |
| 535   | Virgo | Milkyway  | 6/1/1985  | 5/1/1985   | 0.07 | 0.008245 | 0.06 |
| 1056  | Virgo | Andromeda | 6/1/1993  | 8/1/1993   | 1.57 | NaN      | 1.59 |
| 1046  | Virgo | Andromeda | 7/1/1993  | 8/1/1993   | 1.58 | NaN      | 1.59 |
| 956   | Virgo*| Andromeda*| 8/1/1993* | 8/1/1993*  | 1.59*| 0.006688*| 1.59 |
| 776   | Virgo | Andromeda | 9/1/1993  | 8/1/1993   | 1.60 | 0.012917 | 1.59 |

Answer 1

我首先根据条件创建新列，然后填充

df['new']=np.where(df['book_date'] == df['start_date'], df['wf'],np.nan)

df['new'] = df.groupby(['Group', 'SubGroup']).new.apply(lambda x: x.ffill().bfill())


        Group   SubGroup    book_date   start_date  wf      co2         new
236     Virgo   Milkyway    3/1/1985    5/1/1985    0.04    NaN         0.06
239     Virgo   Milkyway    4/1/1985    5/1/1985    0.05    NaN         0.06
1178    Virgo   Milkyway    5/1/1985    5/1/1985    0.06    0.004179    0.06
535     Virgo   Milkyway    6/1/1985    5/1/1985    0.07    0.008245    0.06
1056    Virgo   Andromeda   6/1/1993    8/1/1993    1.57    NaN         1.59
1046    Virgo   Andromeda   7/1/1993    8/1/1993    1.58    NaN         1.59
956     Virgo   Andromeda   8/1/1993    8/1/1993    1.59    0.006688    1.59
776     Virgo   Andromeda   9/1/1993    8/1/1993    1.60    0.012917    1.59

Answer 2

获取您的数据

df = pd.read_clipboard()
df.head()
df.replace({'\*': ''}, regex=True, inplace=True)

def gen_new_col(frame):
    if frame['book_date'] == frame['start_date']:
        return frame['wf']
    else:
        return 'ignore'

df['new_col'] = df.apply(gen_new_col, axis=1)
df['g_subg'] = df['Group'] + "|" + df['SubGroup']

df

    Group   SubGroup    book_date   start_date  wf  co2 new_col g_subg
0   Virgo   Milkyway    3/1/1985    5/1/1985    0.04    NaN ignore  Virgo|Milkyway
1   Virgo   Milkyway    4/1/1985    5/1/1985    0.05    NaN ignore  Virgo|Milkyway
2   Virgo   Milkyway    5/1/1985    5/1/1985    0.06    0.004179    0.06    Virgo|Milkyway
3   Virgo   Milkyway    6/1/1985    5/1/1985    0.07    0.008245    ignore  Virgo|Milkyway
4   Virgo   Andromeda   6/1/1993    8/1/1993    1.57    NaN ignore  Virgo|Andromeda
5   Virgo   Andromeda   7/1/1993    8/1/1993    1.58    NaN ignore  Virgo|Andromeda
6   Virgo   Andromeda   8/1/1993    8/1/1993    1.59    0.006688    1.59    Virgo|Andromeda
7   Virgo   Andromeda   9/1/1993    8/1/1993    1.6 0.012917    ignore  Virgo|Andromeda

# Get a lookup
valid = df[df['new_col'] != 'ignore']
lookup = dict(zip(valid['g_subg'], valid['new_col']))
lookup
{'Virgo|Andromeda': '1.59', 'Virgo|Milkyway': '0.06'}

# Bring it back in
df['final_value']  = df['g_subg'].map(lambda x: lookup[x])
df

    Group   SubGroup    book_date   start_date  wf  co2 new_col g_subg  final_value
0   Virgo   Milkyway    3/1/1985    5/1/1985    0.04    NaN ignore  Virgo|Milkyway  0.06
1   Virgo   Milkyway    4/1/1985    5/1/1985    0.05    NaN ignore  Virgo|Milkyway  0.06
2   Virgo   Milkyway    5/1/1985    5/1/1985    0.06    0.004179    0.06    Virgo|Milkyway  0.06
3   Virgo   Milkyway    6/1/1985    5/1/1985    0.07    0.008245    ignore  Virgo|Milkyway  0.06
4   Virgo   Andromeda   6/1/1993    8/1/1993    1.57    NaN ignore  Virgo|Andromeda 1.59
5   Virgo   Andromeda   7/1/1993    8/1/1993    1.58    NaN ignore  Virgo|Andromeda 1.59
6   Virgo   Andromeda   8/1/1993    8/1/1993    1.59    0.006688    1.59    Virgo|Andromeda 1.59
7   Virgo   Andromeda   9/1/1993    8/1/1993    1.6 0.012917    ignore  Virgo|Andromeda 1.59

将多个值切成单个列

2 个答案: