Pivot Pandas Dataframe 添加列

Question

I have the following dataframe:

date           product   ...   cost   quantity
2018-01-02     orange    ...    7.5       2

2018-01-02     apples    ...    10        5
2018-01-02     apples    ...    12        4

2018-01-04     melon    ...    6.5       10
2018-01-04     melon    ...    5          4
2018-01-04     melon    ...    3.2        3

...

I would like to create the following dataframe, in which each row represent a single date/product combination and in which cost_x and quantity_x are added as additional columns.

More in detail, cost_n and quantity_n represent cost and quantities associated with the last appearing column ( n is an integer);

to illustrate:

date           product   ...   cost_0   quantity_0  cost_1   quantity_1  cost_2   quantity_2 ... cost_ n  quantity_n
2018-01-02     orange    ...    7.5        2          0          0         0         0                0      0
2018-01-02     apples    ...    10         5          12         4         0         0                0      0
2018-01-04     melon    ...    6.5        10          5          4       3.2         3                0      0

How can I create that?

Answer 1

Just modifying user3483203 's answer

x = (df.assign(flag=df.groupby(['date', 'product']).cost.cumcount())
    .pivot_table(index=['date', 'product'], columns='flag', values='cost', aggfunc='first')
    .add_prefix('cost_'))

y = (df.assign(flag=df.groupby(['date', 'product']).cost.cumcount())
    .pivot_table(index=['date', 'product'], columns='flag', values='quantity', aggfunc='first')
    .add_prefix('quantity_'))

z = pd.merge(x,y, on = ['date', 'product']).reset_index()

---

z:

flag    date        product cost_0  cost_1  cost_2  quantity_0  quantity_1  quantity_2
0       2018-01-02  apples  10.0    12.0    NaN     5.0         4.0         NaN
1       2018-01-02  orange  7.5     NaN     NaN     2.0         NaN         NaN
2       2018-01-04  melon   6.5     5.0     3.2     10.0        4.0         3.0