Python 基于列上指定的级别转换数据帧

我想转换这个

DataFrame

，它有一列

Level

，指定行的层次结构级别

df_1 = pd.DataFrame(np.array([["Steel: Furnaces, Refining and Rolling", 1, 155.568345, 152.042158, 152.587873], 
                       ["Steel: Furnaces, Refining and Rolling - Thermal", 2, 99.841607, 97.472990, 97.822843],
                       ["LPG", 3, 0.300934, 0.000000, 0.000000], 
                       ["Diesel oil (incl. biofuels)", 3, 0.000000, 0.000000, 0.000000], 
                       ["Residual fuel oil", 3, 32.204198, 31.535245, 31.648432], 
                       ["Natural gas (incl. biogas)", 3, 67.336475, 65.937745, 66.174411], 
                       ["Steel: Furnaces, Refining and Rolling - Electric", 2, 55.726738, 54.569168, 54.765030]]), 
                       columns=['AT: Iron and steel / useful energy demand', 'level', '2000', '2001', '2002'])
df_1 

+--------------------------------------------------+-------+------------+------------+------------+
|    AT: Iron and steel / useful energy demand     | level |    2000    |    2001    |    2002    |
+--------------------------------------------------+-------+------------+------------+------------+
| Steel: Furnaces, Refining and Rolling            |     1 | 155.568345 | 152.042158 | 152.587873 |
| Steel: Furnaces, Refining and Rolling - Thermal  |     2 |  99.841607 |   97.47299 |  97.822843 |
| LPG                                              |     3 |   0.300934 |        0.0 |        0.0 |
| Diesel oil (incl. biofuels)                      |     3 |        0.0 |        0.0 |        0.0 |
| Residual fuel oil                                |     3 |  32.204198 |  31.535245 |  31.648432 |
| Natural gas (incl. biogas)                       |     3 |  67.336475 |  65.937745 |  66.174411 |
| Steel: Furnaces, Refining and Rolling - Electric |     2 |  55.726738 |  54.569168 |   54.76503 |
+--------------------------------------------------+-------+------------+------------+------------+

像这样的事情。一级是过程，二级是能量，三级是燃料

df_2 = pd.DataFrame({'Process': ["Steel: Furnaces, Refining and Rolling", "Steel: Furnaces, Refining and Rolling", "Steel: Furnaces, Refining and Rolling", "Steel: Furnaces, Refining and Rolling", "Steel: Furnaces, Refining and Rolling", "Steel: Furnaces, Refining and Rolling", "Steel: Furnaces, Refining and Rolling"], 
                     'Energy': [None, 'Steel: Furnaces, Refining and Rolling - Thermal', 'Steel: Furnaces, Refining and Rolling - Thermal', 'Steel: Furnaces, Refining and Rolling - Thermal', 'Steel: Furnaces, Refining and Rolling - Thermal', 'Steel: Furnaces, Refining and Rolling - Thermal', 'Steel: Furnaces, Refining and Rolling - Electric'],
                     'Fuel':  [None, None, 'LPG', 'Diesel oil (incl. biofuels)', 'Residual fuel oil', 'Natural gas (incl. biogas)', None], 
                     '2000':  [155.5683448, 99.84160689, 0.300933684, 0, 32.20419829, 67.33647492, 55.72673787], 
                     '2001':  [152.0421582, 97.47298987, 0, 0, 31.53524476, 65.93774511, 54.56916837], 
                     '2002':  [152.5878732, 97.82284329, 0, 0, 31.64843215, 66.17441114, 54.76502991]}) 
df_2

+---------------------------------------+--------------------------------------------------+-----------------------------+------------+------------+------------+
| Process                               | Energy                                           | Fuel                        | 2000       | 2001       | 2002       |
+---------------------------------------+--------------------------------------------------+-----------------------------+------------+------------+------------+
| Steel: Furnaces, Refining and Rolling | None                                             | None                        | 155.568345 | 152.042158 | 152.587873 |
| Steel: Furnaces, Refining and Rolling | Steel: Furnaces, Refining and Rolling - Thermal  | None                        | 99.841607  | 97.472990  | 97.822843  |
| Steel: Furnaces, Refining and Rolling | Steel: Furnaces, Refining and Rolling - Thermal  | LPG                         | 0.300934   | 0.000000   | 0.000000   |
| Steel: Furnaces, Refining and Rolling | Steel: Furnaces, Refining and Rolling - Thermal  | Diesel oil (incl. biofuels) | 0.000000   | 0.000000   | 0.000000   |
| Steel: Furnaces, Refining and Rolling | Steel: Furnaces, Refining and Rolling - Thermal  | Residual fuel oil           | 32.204198  | 31.535245  | 31.648432  |
| Steel: Furnaces, Refining and Rolling | Steel: Furnaces, Refining and Rolling - Thermal  | Natural gas (incl. biogas)  | 67.336475  | 65.937745  | 66.174411  |
| Steel: Furnaces, Refining and Rolling | Steel: Furnaces, Refining and Rolling - Electric | None                        | 55.726738  | 54.569168  | 54.765030  |
+---------------------------------------+--------------------------------------------------+-----------------------------+------------+------------+------------+

我如何才能做到这一点？

使用-

levs = df_1.loc[df_1['level']!='1', 'level'].unique()
ind = df_1['AT: Iron and steel / useful energy demand'].where(df_1['level']=='1').ffill().values
for j, i in enumerate(levs):
    lev = df_1['AT: Iron and steel / useful energy demand'].where(df_1['level']==i)
    if j != (len(levs)-1):
        lev = lev.ffill()
    ind = np.vstack((ind, lev))
ind = pd.MultiIndex.from_tuples(list(zip(*ind)))
tmp = df_1.set_index(ind).reset_index()
tmp = tmp.rename(columns={'level_0': 'Process', 'level_1':'Energy', 'level_2':'Fuel'}).drop(['level'], axis=1)
tmp

这样做：

# unstack the column "level" 
df_3 = df_1.set_index(['level', '2000', '2001', '2002'], append=True).unstack('level')
df_3.columns = df_3.columns.droplevel(0)

# forward fill (propagate last valid observation forward)
df_3[['1', '2']] = df_3[['1', '2']].ffill()

df_3 = df_3.reset_index().drop(['level_0'], axis=1)
df_3 = df_3.rename(columns={'1': 'Process', '2':'Energy', '3':'Fuel'})
column_names = ['Process', 'Energy', 'Fuel'] + ['2000', '2001', '2002']
df_3 = df_3.reindex(columns=column_names)

df_3

df_1.设置索引（['level'、'2000'、'2001'、'2002']，append=True）。取消堆栈（“level”）

获取所需格式，然后

ffill

逐列显示您的条件。