我正在制作一个程序，根据其排名为供应商分配数量。数据是这样的：
Origin Dest Provider Vol_A Vol_B Vol_C Capacity rank
NYC AMS A 90 1300 2500 4000 1
NYC AMS B 150 600 1700 3000 2
NYC BRI A 105 700 100 2300 1
NYC BRI C 300 1300 200 2800 2
目标是将所有卷分配给排名第一的提供商，直到满足Capacity为止。在第一个例子中，NYC到AMS，提供者A将被分配240个单位Vol_A，因为90 + 150 = 240.所需的输出如下所示：
Origin Dest Provider Vol_A Vol_B Vol_C Capacity rank assg_a assg_b assg_c
NYC AMS A 90 1300 2500 4000 1 240 1900 1980
NYC AMS B 150 600 1700 3000 2 0 0 2220
NYC BRI A 105 700 100 2300 1 405 1895 0
NYC BRI C 300 1300 200 2800 2 0 105 300
在NYC-AMS示例中，提供商A无法填写所有Vol_C，因此其中一些溢出到提供商B.
我的代码如下：
def assign():
vol_sim = pd.DataFrame(columns=['Origin','Dest','Provier','rank','Vol_E','Vol_C','Vol_U','Capacity','assg_e','assg_c','assg_u'])
for key,lane in inputs.groupby(['Origin','Dest'])
for col,out in zip(['Vol_E','Vol_C','Vol_U'],['assg_e','assg_c','assg_u']):
to_assg = lane[col].sum()
assg = 0
remain = to_assg
for idx,row in lane.iterrows():
if assg >= to_assg:
row[out] = 0
row_temp = pd.DataFrame(row[['Origin','Dest','Provider','rank',col,'Capacity',out]])
row_temp2 = row_temp.T
vol_sim = vol_sim.append(row_temp2)
else:
if row['Capacity'] <= to_assg:
row[out] = row['Capacity']
assg = assg + row[out]
remain = remain - assg
row_temp = pd.DataFrame(row[['Origin','Dest','Provider','rank',col,'Capacity',out]])
row_temp2 = row_temp.T
vol_sim = vol_sim.append(row_temp2)
else:
row[out] = remain
assg = assg + row[out]
remain = remain - assg
row_temp = pd.DataFrame(row[['Origin','Dest','Provider','rank',col,'Capacity',out]])
row_temp2 = row_temp.T
vol_sim = vol_sim.append(row_temp2)
return vol_sim
我得到的结果（如下所示）似乎在列上的每次迭代都有重复。我希望我的结果与我想要的输出格式相同，而不是nans。我想我可以通过分组数据来做到这一点，但我宁愿在函数本身内完成它。有什么想法吗？
Origin Dest Provider Vol_A Vol_B Vol_C Capacity rank assg_a assg_b assg_c
NYC AMS A 90 1300 2500 4000 1 120 nan nan
NYC AMS B 150 600 1700 3000 2 0 nan nan
NYC AMS A 90 1300 2500 4000 1 nan 1900 nan
NYC AMS B 150 600 1700 3000 2 nan 0 nan
NYC AMS A 90 1300 2500 4000 1 nan nan 1980
NYC AMS B 150 600 1700 3000 2 nan nan 2220