如何在Python中汇总字典并按键值排序

Question

I have a log file of net-flow data that I am trying to sort by ip address and time stamp and add the bytes. Thus, it needs to list the same ip address in descending order by byte amount.

The output of the file reads:

                 Min       Source IP                     Bytes

./R2snd/2014/02/02/02/25.flows:100.000.000.000|101.101.101.101|0|4|3|2|96|1391336665|1391336668|3361|445|2|6|0|0|0|0|0

For some reason I can only get it to display the minute but i need the whole time and date formatted. The minute is the last /number i typed minute above it. Then I need it to take every ip address in the file and sort them by ip thus repeating ips would appear together, and add the bytes send for each ip. I have tried to do this below with a dictionary but I can't seem to get it to work. Then I need to sort the dictionary in descending order by bytes, because for each ip entry it needs to add the bytes, thus the top entry for each ip will be the total bytes sent by that ip.

import operator
with open('/home/username/Documents/log') as f:
    for line in f:
        #save the data into an array
        firstsplitforminute = line.split('/')
        secondsplitforminute = firstsplitforminute[6].split('.')
        firstsplitforsourceip = line.split('|')
        secondsplitforsourceip = firstsplitforsourceip[0].split(':')
        minute = secondsplitforminute[0]
        sourceip = secondsplitforsourceip[1]
        bytes = line.split('|')[6]
        protocol = line.split('|')[12]

        if protocol == '6':
            entries = {'IP':sourceip, 'BYTES':bytes, 'MIN':minute}
            sum(item['BYTES'] for item in entries)
            def sortbykey():
                sortedbykeydict = sorted(entries.items(), key = lambda t: t[1])
                print sortedbykeydict
             sortbykey() 
        else:
            pass

however I get the following error when I run this code:

File "/home/grant/.eclipse/org.eclipse.platform_3.8_155965261/plugins/org.python.pydev_3.4.1.201403181715/pysrc/pydevd.py", line 1844, in <module>
    debugger.run(setup['file'], None, None)
  File "/home/grant/.eclipse/org.eclipse.platform_3.8_155965261/plugins/org.python.pydev_3.4.1.201403181715/pysrc/pydevd.py", line 1372, in run
    pydev_imports.execfile(file, globals, locals)  # execute the script
  File "/home/grant/workspace/Learning/LogfileExtractor.py", line 16, in <module>
    sum(item['BYTES'] for item in entries)
  File "/home/grant/workspace/Learning/LogfileExtractor.py", line 16, in <genexpr>
    sum(item['BYTES'] for item in entries)
TypeError: string indices must be integers, not str

Answer 1

Try parsing: 'BYTES':int(bytes)

(For as far as I understand your code that should work)

Answer 2

@BartoszKP is correct. Python is looping through entries , which will not result in a string:

entries = {'IP':sourceip, 'BYTES':bytes, 'MIN':minute}
sum(item['BYTES'] for item in entries)

Instead, you should "itemize" the dictionary:

sum(v for k,v in entries.items())

This means that during the first iteration 'IP' is stored in k and sourceip is stored in v ; the second, 'BYTES' is stored in k and bytes is stored in v ; and so on...