Python 数据框 - 如何 evaluate/use 动态创建列

Question

假设我有一个数据框如下：

+---------+-------+------------+
| Product | Price | Calculated |
+---------+-------+------------+
| A       |    10 | 10         |
| B       |    20 | NaN        |
| C       |    25 | NaN        |
| D       |    30 | NaN        |
+---------+-------+------------+

以上可以使用下面的代码创建：

data = {'Product':['A', 'B', 'C', 'D'],
        'Price':[10, 20, 25, 30],
        'Calculated':[10, np.nan, np.nan, np.nan]}

df = pd.DataFrame(data)

我想即时更新列 calculated。对于第二行，第 2 行的 calculated = Prv. calculated / Previous Price 即 calculated 是 10/10=1

现在我们有了第 2 行 calculated 的值，第 3 行 calculated 将是 1/20 等等。

预期输出

+---------+-------+------------+
| Product | Price | Calculated |
+---------+-------+------------+
| A       |    10 |         10 |
| B       |    20 |          1 |
| C       |    25 |       0.05 |
| D       |    30 |      0.002 |
+---------+-------+------------+

以上可以使用循环来实现，但我不想使用循环，而是需要一种矢量化方法来更新列 Calculated。我怎样才能做到这一点？

Answer 1

您正在查看 cumprod shift:

# also `df['Calculated'].iloc[0]` instead of `.ffill()`
df['Calculated'] = df['Calculated'].ffill()/df.Price.cumprod().shift(fill_value=1)

输出：

  Product  Price  Calculated
0       A     10      10.000
1       B     20       1.000
2       C     25       0.050
3       D     30       0.002

Python 数据框 - 如何 evaluate/use 动态创建列

Python Data Frame - How can I evaluate/use a column being created on the fly

pandas

dataframe

vectorization

fillna