[英]Pandas GroupBy: comma separated list of sums
I have the below groupby
which is summing the Amounts at the "ParentAccount" level.我有以下
groupby
,它在“ParentAccount”级别对金额求和。 I am trying to show on the same line the details behind that sum.我试图在同一行上显示该金额背后的详细信息。 I have the comma separated list of accounts showing next to the amount total but would also like to add in a single column that shows a comma separated sum at the account level.
我在总金额旁边显示了逗号分隔的帐户列表,但也想添加一个单独的列,在帐户级别显示逗号分隔的总和。
So for the below code I would have the following float strings in a separate column所以对于下面的代码,我将在单独的列中包含以下浮点字符串
ParentAccount 1: 3.75, 1
ParentAccount 2: 14, 10.5
Not sure of the best way to go about doing this.不确定有关执行此操作的最佳方法 go。 I tried doing a merge of two separate
groupby
s but think there is probably a better way of doing this.我尝试合并两个单独的
groupby
,但认为可能有更好的方法。
import pandas as pd
data = {
'ParentAccount': [1,1,1,2,2,2],
'Account': ['A', 'A', 'C', 'D', 'D','E'],
'Amount': [1.5, 2.25, 1, 4.75, 9.25, 10.50],
}
df = pd.DataFrame(data)
df_final = df.groupby('ParentAccount').agg({'Amount': 'sum', 'Account': lambda x: ','.join(x.unique()),}).add_suffix('-Net')
print(df_final)
You could groupby
"ParentAccount" and "Account" to find the sum
;您可以按“
groupby
”和“Account”分组来查找sum
; then groupby
"ParentAccount" again, and pass an unpacked dictionary to agg
to do the things you want: (i) Summing the amount and (ii) join
ing the unique accounts for each ParentAccount (iii) join
ing the amounts per account for each ParentAccount:然后
groupby
"ParentAccount" 再次,并将解压缩的字典传递给agg
来做你想做的事情:(i)总结金额和(ii) join
每个ParentAccount的唯一账户(iii) join
每个账户的金额家长帐户:
out = (df
.groupby(['ParentAccount','Account'])
.sum()
.reset_index(level=1)
.groupby(level=0)
.agg(**{'Amount-Net': ('Amount','sum'),
'Account-Net': ('Account', lambda x: ', '.join(x)) ,
'Amounts per Account': ('Amount', lambda x: ', '.join(x.astype(str)))}))
Output: Output:
Amount-Net Account-Net Amounts per Account
ParentAccount
1 4.75 A, C 3.75, 1.0
2 24.50 D, E 14.0, 10.5
Use a double groupby
:使用双
groupby
:
out = (
df.groupby(['ParentAccount', 'Account'], as_index=False)['Amount'].sum()
.groupby('ParentAccount', as_index=False)
.agg(**{'Amount-Net': ('Amount', 'sum'),
'Amount-Detail': ('Amount', lambda x: ','.join(x.astype(str))),
'Account-Net': ('Account', ','.join)})
)
Output: Output:
>>> out
ParentAccount Amount-Net Amount-Detail Account-Net
0 1 4.75 3.75,1.0 A,C
1 2 24.50 14.0,10.5 D,E
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.