pandas 数据框-如何将多行合并为一个 [英] Pandas Dataframe - How to combine multiple rows to one

问题描述

我有一个数据集，格式为:

I have a dataset in the form:

    A            B
0  30  60538815980
1  30   7410811099
2  26   2238403510
3  26   2006613744
4  26   2006618783
5  26   2006613743

我想合并A值匹配的行并产生类似的结果

I want to combine the rows where the value of A matches and produce something like that

            C_1         C_2         C_3         C_4
A                                                  
26   2238403510  2006613744  2006618783  2006613743
30  60538815980  7410811099         NaN         NaN

我尝试用连接或合并来表达它，但到目前为止却失败了.有什么简单的方法可以表达这一点，还是我必须使用apply并创建一个新的DataFrame?

I have tried expressing it in terms of join or merge but have failed so far. Is there any simple way to express that or will I have to use apply and create a new DataFrame?

推荐答案

首先，基于列A创建一个groupby对象.然后创建一个新的数据框df2，该数据框使用ix根据列A中的值n为每个组的列B编制索引.将此数据帧的索引设置为等于groupby中的键值(即列A中的唯一值).

First, create a groupby object based on column A. Then create a new dataframe df2 which uses ix to index column B of each group based on the value n from column A. Set the index of this dataframe equal to the key values from the groupby (i.e. the unique values from column A).

最后，使用列表推导将新的列值设置为等于C_1，C_2，...等.

Finally, use a list comprehension to set the new column values equal to C_1, C_2, ..., etc.

df = pd.DataFrame({'A': [30, 30, 26, 26, 26, 26], 
                   'B': [60538815980, 7410811099, 2238403510, 
                         2006613744, 2006618783, 2006613743]})

gb = df.groupby('A')
df2 = pd.DataFrame([df.ix[gb.groups[n], 'B'].values for n in gb.groups], 
                   index=gb.groups.keys())
df2.columns = ["C_" + str(i + 1) for i in df2.columns]
df2.index.name = "A"

>>> df2
            C_1         C_2         C_3         C_4
A
26   2238403510  2006613744  2006618783  2006613743
30  60538815980  7410811099         NaN         NaN

这篇关于 pandas 数据框-如何将多行合并为一个的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！