数据

dict = {
    'Deal' : [1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 4, 4, 4, 4],
    'Flow' : ['Investment', 'Borrow', 'Interest', 'Capital', 'Investment', 'Borrow', 'Interest', 'Capital', 'Investment',
           'Borrow', 'Interest', 'Investment', 'Borrow', 'Interest', 'Capital'],
    'Amount' : [100, 10 ,5 ,50 ,100 ,10 ,5 ,50 ,100 ,10 ,5 ,100 ,10 ,5 ,50 ],
    'Currency' : ['USD' ,'USD' ,'USD' ,'EUR' ,'USD' ,'USD' ,'USD' ,'USD' ,'USD' ,'EUR' ,'USD' ,'USD' ,'EUR' ,'USD' ,'USD']
}

df = pd.DataFrame(dict)

创建新列以区分基于流量的负金额。

df['New_Amount'] = df.apply(lambda x : -1 * x['Amount'] if x['Flow'] in ['Borrow', 'Interest'] else x['Amount'], axis=1)

创建交易列表，其中流计数为4，然后筛选 Dataframe

dealed_df = df.groupby('Deal').agg({'Flow':'count'}).reset_index()
list_deal = dealed_df[dealed_df['Flow']==4]['Deal'].to_list()
filtered_df = df[df['Deal'].isin(list_deal)]

然后根据交易按 Dataframe 分组，对new_amount求和

filtered_df.groupby(['Deal', 'Currency']).agg({'New_Amount':'sum'}).reset_index().pivot(index='Deal', columns='Currency', values='New_Amount').reset_index()

结果

Currency    Deal    EUR     USD
0           1       50.0    85.0
1           2       NaN     135.0
2           4       -10.0   145.0