按多个列分组并在pandas中生成格式化结果

Question

这是我的文字文件

No.,Time,Source,Destination,Protocol,Length,Info,SrcPort,DstPort,src_dst_pair
1401,0.397114,145.95.225.186,210.218.218.164,UDP,100,Source port: hsrp  Destination port: hsrp,hsrp,1985,"('145.95.225.186', '210.218.218.164')"
8999,3.229111,145.95.225.186,210.218.218.164,UDP,100,Source port: hsrp  Destination port: hsrp,hsrp,1985,"('145.95.225.186', '210.218.218.164')"
18504,5.877098,145.95.225.186,210.218.218.164,UDP,100,Source port: hsrp  Destination port: hsrp,hsrp,1985,"('145.95.225.186', '210.218.218.164')"
23755,8.695843,145.95.225.186,210.218.218.164,UDP,100,Source port: hsrp  Destination port: hsrp,hsrp,1985,"('145.95.225.186', '210.218.218.164')"
28027,11.24121,145.95.225.186,210.218.218.164,UDP,100,Source port: hsrp  Destination port: hsrp,hsrp,1985,"('145.95.225.186', '210.218.218.164')"
33304,14.117213,145.95.225.186,210.218.218.164,UDP,100,Source port: hsrp  Destination port: hsrp,hsrp,1985,"('145.95.225.186', '210.218.218.164')"
700443,222.305789,145.95.41.251,145.95.81.118,UDP,50,Source port: 36477  Destination port: snmp,36477,161,"('145.95.41.251', '145.95.81.118')"
700495,222.351933,145.95.41.251,145.95.81.118,UDP,50,Source port: 36477  Destination port: snmp,36477,161,"('145.95.41.251', '145.95.81.118')"
700496,222.352372,145.95.41.251,145.95.81.118,UDP,50,Source port: 36477  Destination port: snmp,36477,161,"('145.95.41.251', '145.95.81.118')"
708982,225.913385,145.95.41.251,145.95.81.118,UDP,50,Source port: 36477  Destination port: snmp,36477,161,"('145.95.41.251', '145.95.81.118')"
709797,226.130847,145.95.41.251,145.95.81.118,UDP,50,Source port: 36477  Destination port: snmp,36477,161,"('145.95.41.251', '145.95.81.118')"
710340,226.372421,145.95.41.251,145.95.81.118,UDP,50,Source port: 36477  Destination port: snmp,36477,161,"('145.95.41.251', '145.95.81.118')"

我想基于Source和Destination对数据进行分组，然后：

1. 在组

中累积长度列

2. 在组中找到最大和最小时间之间的差异

我得到了结果，但我需要按照我在预期输出中显示的方式对其进行格式化。我也想知道是否有更好的方法来做到这一点。

以下是我的尝试

import pandas as pd

data = pd.read_csv('simple_udp.csv')
# getting the accumulated sum for the group
length = data.groupby(['Source','Destination']).Length.sum()
# getting the difference in time between the max and min in the group
time  = data.groupby(['Source','Destination']).Time.max() - data.groupby(['Source','Destination']).Time.min()
# This is were I have problem. How can i format the result so that 
# I can get the expected output(shown below) 
print length, time

预期输出

Source          Destination       Length  Time
145.95.225.186  210.218.218.164    600    13.720099
145.95.41.251   145.95.81.118      300     4.066632

Answer 1

使用agg

data.groupby(['Source','Destination']).agg({'Length': 'sum', 'Time': lambda x: x.max() - x.min()})

Answer 2

我的第一个猜测是

import pandas as pd
data = pd.read_csv('simple_udp.csv')
# Creating a DataFramGroupBy object
group = data.groupby(['Source','Destination'])
df_length = g['Length'].sum()
df_time   = g['Time'].max() - g['Time'].min()
df = pd.DataFrame([df_length,df_time])

或者如果您希望在较少的行上使用它，但也不太可读，请使用agg上的group方法