DGMS89's questions

场景：我有一个包含一行数据的数据框。每列代表一年，并具有相关值。我尝试使用此 df 中的数据创建一个 json 样式结构以传递给 API 请求。post。

样本 DF：

+-------+-------+-------+-------+-------+-------+-------------+-------------+-------------+-------------+-------------+-------------+
|       | 2020  | 2021  | 2022  | 2023  | 2024  |    2025     |    2026     |    2027     |    2028     |    2029     |    2030     |
+-------+-------+-------+-------+-------+-------+-------------+-------------+-------------+-------------+-------------+-------------+
| Total | 23648 | 20062 | 20555 | 22037 | 26208 | 28224.88801 | 29975.87934 | 31049.01582 | 32170.68853 | 33190.35298 | 34031.93951 |
+-------+-------+-------+-------+-------+-------+-------------+-------------+-------------+-------------+-------------+-------------+

JSON 样式结构示例：

parameters = {
    "first_Id":first_id,
    "version":2,
    "overrideData":[
      {
          "period":2024,          
          "TOTAL":101.64,          
      },
      {
          "period":2025,
          "TOTAL":104.20,
      }
    ]
}

问题：使用 Df 中的数据填充和扩展 JSON 样式对象的最佳方法是什么？我尝试了以下方法，但这只会分隔两行，一行用于总计，一行用于期间，这不会导致配对需求：

        parameters = {}
        parameters['first_Id'] = first_id
        parameters['version'] = 2
        parameters['overrideData'] = {
            }
        parameters['overrideData']['total'] = test_input.iloc[0].tolist()         
        parameters['overrideData']['period'] = list(test_input.columns)

其结果是：

{
  "companyId": 11475,
  "version": 2,
  "overrideData": {
    "TOTAL": [
      23647.999999999996,
      20061.999999999996,
      20555,
      22036.999999999996,
      26207.999999999993,
      28224.88800768,
      29975.879336500002,
      31049.015816740008,
      32170.68852577,
      33190.3529754,
      34031.93951397
    ],
    "period": [
      2020,
      2021,
      2022,
      2023,
      2024,
      2025,
      2026,
      2027,
      2028,
      2029,
      2030
    ]
  }
}

场景：对于具有重复索引的数据框，我想在保留最大值的同时进行分组。我在按索引删除重复项，在重复项中保留每列的最大值中找到了解决方案，但是，这会获取每列的最大值。这会混合不同行的数据，保留最大值。

问题：如果我不想混合不同行的值，而是想要保留一行，其中“C”列的值在具有相同索引的行中最高（在这种情况下，我将选择“C”中值最高的行并保留该行的所有值，不与其他行中其他列的高值混合），应该如何执行 groupby？

我尝试过：从链接的问题中，我得到了

df.groupby(df.index).max()

并尝试将其修改为：

df.groupby(df.index)['C'].max()

但这会删除数据框的其他列。

场景：我有一个数据框，我想根据其他列的值对某一列内的行求和。

DF示例：

+----------+----------------------------+---------------------+---------+
| SCENARIO |          PRODUCT           |        FLOW         |  2030   |
+----------+----------------------------+---------------------+---------+
| Stated   | Natural gas: unabated      | Total energy supply | 147.541 |
| Stated   | Natural gas: with CCUS     | Power sector inputs |   0.018 |
| Stated   | Natural gas: with CCUS     | Total energy supply |   1.147 |
| Stated   | Nuclear                    | Power sector inputs |  36.773 |
| Stated   | Nuclear                    | Total energy supply |  36.773 |
| Stated   | Oil                        | Power sector inputs |   5.162 |
| Stated   | Oil                        | Total energy supply | 194.938 |
| Stated   | Renewables                 | Power sector inputs |  76.612 |
| Stated   | Renewables                 | Total energy supply | 119.979 |
| Stated   | Solar                      | Total energy supply |   22.66 |
| Stated   | Solar PV                   | Power sector inputs |  19.457 |
| Stated   | Total                      | Power sector inputs |   263.2 |
| Stated   | Total                      | Total energy supply | 667.905 |
| Stated   | Traditional use of biomass | Total energy supply |  18.975 |
| Stated   | Wind                       | Power sector inputs |  18.823 |
| Stated   | Wind                       | Total energy supply |  18.823 |
+----------+----------------------------+---------------------+---------+

想要的结果：我想根据其他列的条件对给定列（在本例中为 2030）的行值求和。例如：当场景 = 状态、流量 = 总能源供应时，我想对产品为风能和太阳能的 2030 年的值求和，因此在本例中，我的结果将是一个新行，其中 2030 列的值为 22.66（太阳能值）+ 18.823（风能值），即 41.483。它看起来像：

+----------+----------------------------+---------------------+---------+
| SCENARIO |          PRODUCT           |        FLOW         |  2030   |
+----------+----------------------------+---------------------+---------+
| Stated   | Natural gas: unabated      | Total energy supply | 147.541 |
| Stated   | Natural gas: with CCUS     | Power sector inputs |   0.018 |
| Stated   | Natural gas: with CCUS     | Total energy supply |   1.147 |
| Stated   | Nuclear                    | Power sector inputs |  36.773 |
| Stated   | Nuclear                    | Total energy supply |  36.773 |
| Stated   | Oil                        | Power sector inputs |   5.162 |
| Stated   | Oil                        | Total energy supply | 194.938 |
| Stated   | Renewables                 | Power sector inputs |  76.612 |
| Stated   | Renewables                 | Total energy supply | 119.979 |
| Stated   | Solar                      | Total energy supply |   22.66 |
| Stated   | Solar PV                   | Power sector inputs |  19.457 |
| Stated   | Total                      | Power sector inputs |   263.2 |
| Stated   | Total                      | Total energy supply | 667.905 |
| Stated   | Traditional use of biomass | Total energy supply |  18.975 |
| Stated   | Wind                       | Power sector inputs |  18.823 |
| Stated   | Wind                       | Total energy supply |  18.823 |
| Stated   | result_sum                 |                     |  41.483 |
+----------+----------------------------+---------------------+---------+

问题：尝试使用 df.sum 函数的不同组合时，我无法获得预期的结果。如果我尝试使用运算符（例如 ^ 或 |），例如：

test3 = test3.loc[(test3['SCENARIO'] == 'Stated') & (test3['FLOW'] == 'Total energy supply') & (test3['PRODUCT'] == 'Solar' ^ 'Wind')]

我收到不支持的类型错误：

TypeError: unsupported operand type(s) for |: 'str' and 'str'

如果我尝试使用一系列条件，我不确定如何对必要的 FLOW 值进行多项选择：

test3['sigcontr_steps'] = test3.loc[(test3['SCENARIO'] == 'Stated') & (test3['FLOW'] == 'Total energy supply') & 
                                    (test3['PRODUCT'] == 'Solar') & (test3['PRODUCT'] == 'Wind'),[2022]].sum(axis=0)

由于 PRODUCT 同时具有两个值，因此此方法不起作用。

问题：执行此操作的正确方法是什么？