带有多级索引的熊猫锯齿列

Question

这是我从csv文件读取数据时的样子，我正在使用多级索引（主体类型和支出）进行读取。

我想要有一个“年份”列，“支出”中的所有值都应显示为单独的列。基本上我想将“支出”与“年份”交换/转置

最终数据应如下所示

我找到了一种方法来执行此操作，但是它似乎并不高效。我想知道是否有更好，更清洁的方法？我看不到pd.swapaxes（）的几个示例，但无法正常工作。

这是我使用的代码：

d = [
    ["Small Narrowbodies", "TotalExpenses", "2326550.00", "2566989.00", "2710156.00"],
    ["Small Narrowbodies", "Pilots (000)", "583404.00", "627762.00", "669258.00"],
    [
        "Small Narrowbodies",
        "Salaries and Wages (000)",
        "432613.00",
        "469059.00",
        "515538.00",
    ],
    ["Small Narrowbodies", "Pilot Training (000)", "28235.00", "22388.00", "23838.00"],
    [
        "Small Narrowbodies",
        "Benefits and Payroll Taxes (000)",
        "77752.00",
        "87128.00",
        "77679.00",
    ],
    [
        "Small Narrowbodies",
        "Per Diem/ Personnel (000)",
        "44804.00",
        "49187.00",
        "52203.00",
    ],
    [
        "Small Narrowbodies",
        "Purchased Goods (000)",
        "627471.00",
        "792582.00",
        "772448.00",
    ],
    ["Small Narrowbodies", "Fuel/Oil (000)", "559698.00", "684007.00", "670673.00"],
    ["Small Narrowbodies", "Insurance (000)", "7483.00", "5449.00", "4200.00"],
    [
        "Small Narrowbodies",
        "Other (inc. Tax) (000)",
        "60290.00",
        "103126.00",
        "97575.00",
    ],
]

df = pd.DataFrame(d, columns=["Body_Type", "Spending", "1995", "1996", "1997"])

df2 = df.set_index(["Body_Type", "Spending"])

df3 = df2.transpose().unstack(level=-1).reset_index()

df3.columns = ["Body_Type", "Spending", "Year", "Amount"]

df4 = df3.pivot_table(
    index["Body_Type", "Year"], columns="Spending", values="Amount", aggfunc=np.sum)

Answer 1

这更像是

df=df.unstack(level=0).stack(level=0)