Question

我有一个类似下面的df，我想在一个新的DataFrame中对其进行转换

# column_1  column_2    column_3  column_4
# ticket    12345     
# Date      2020-02-01
# UPC Code  Description Qty       Unit Price
# 987654    product 1   1         10
# 879756    product 2   1         7
# 987895    product 3   2         5
# ticket    12346
# Date      2020-02-03
# UPC Code  Description Qty       Unit Price
# 987654    product 1   1         10
# 997651    product 4   1         3 
# ticket    12347

这是新数据框的示例：

# ticket  date        upc_code  description qty unit_price
# 12345   2020-02-01  987654    product 1   1   10
# 12345   2020-02-01  879756    product 2   1   7
# 12345   2020-02-01  987895    product 3   2   5
# 12346   2020-02-03  987654    product 1   1   10
# 12346   2020-02-03  997651    product 4   1   3
# 12347

有人可以帮我吗？我正在尝试找出方法。每个票证值都是一个采购订单，票证和日期值应根据每个订单上的产品重复。然后，Upc代码下面的行随所购买商品的数量而变化。

提前谢谢！

Answer 1

IIUC，您需要按字符串过滤空白行，然后在进行一些旋转后重新连接。

如果空格不是真空值，则可以使用以下代码行将其转换。

import numpy as np
df = df.replace('',np.nan,regex=True)



print(df)

    column_1     column_2 column_3    column_4
0     ticket        12345      NaN         NaN
1       Date   2020-02-01      NaN         NaN
2   UPC Code  Description      Qty  Unit Price
3     987654    product 1        1          10
4     879756    product 2        1           7
5     987895    product 3        2           5
6     ticket        12346      NaN         NaN
7       Date   2020-02-03      NaN         NaN
8   UPC Code  Description      Qty  Unit Price
9     987654    product 1        1          10
10    997651    product 4        1           3
11    ticket        12347      NaN         NaN

s = df.dropna(how='any').loc[~df["column_1"].str.contains("UPC Code")]

s1 = df[~df.index.isin(s.index) & ~df["column_1"].str.contains("UPC Code")]


df2 = pd.concat(
    [
        pd.crosstab(s1.index, s1["column_1"], s1["column_2"], aggfunc="first")
        .ffill()
        .bfill()
        .reset_index(drop=True),
        s.reset_index(drop=True),
    ],
    axis=1,
)

print(df2)

         Date ticket column_1   column_2 column_3 column_4
0  2020-02-01  12345   987654  product 1        1       10
1  2020-02-01  12345   879756  product 2        1        7
2  2020-02-01  12346   987895  product 3        2        5
3  2020-02-03  12346   987654  product 1        1       10
4  2020-02-03  12347   997651  product 4        1        3

将非结构化的熊猫数据框修改为新的结构化数据框

1 个答案: