Python Dataframe：转置Dataframe中的45列并保留日期列而不转置

发布于11月29日

我有一个输入框示例:

Name	Date	Score	Target	Difference
Jim	2023-10-09	9	12	3
Jim	2023-10-16	13	16	3
Andy	2023-10-09	7	7	0
Andy	2023-10-16	5	20	15

用于创建表的Python代码:

import pandas as pd
df = pd.DataFrame({'Name':["Jim","Jim","Andy", "Andy"], 'Date':['2023-10-09', '2023-10-16', '2023-10-09', "2023-10-16"], 'Score':["9","13","7", "5"], 'Target':["12","16","7", "20"], 'Difference':["3","3","0", "15"]})

我想按名称调换上表，并使行为Date、Score、Target和Difference.所需的输出表如下:

Date	Category	Jim	Andy
2023-10-09	Score	9	7
	Target	12	7
	Difference	3	0
2023-10-16	Score	13	5
	Target	16	20
	Difference	3	15

我try 使用下面的代码这样做，但它没有产生所需的按日期和类别列进行转置的表分组.

df_2 =df.T

Df_2生成以下输出，该输出颠倒了我不想要的"Date"列.

推荐答案

您可以先将"Score", "Target", "Difference"列合并为两个键值列，然后应用pivot:

df \
    .melt(
        id_vars=["Name", "Date"],
        value_vars=["Score", "Target", "Difference"],
        var_name="Category"
    ) \
    .pivot(
        index=["Date", "Category"],
        values="value",
        columns="Name"
    )

输出:

Name                  Andy Jim
Date       Category           
2023-10-09 Difference    0   3
           Score         7   9
           Target        7  12
2023-10-16 Difference   15   3
           Score         5  13
           Target       20  16