how to sort a list by the nth element in v2.3?

This is a simple script I wrote:

#!/usr/bin/env python

file = open('readFile.txt', 'r')
lines = file.readlines()
file.close()
del file

sortedList = sorted(lines, key=lambda lines: lines.split('\t')[-2])

file = open('outfile.txt', 'w')
for line in sortedList:
    file.write(line)

file.close()
del file

to rewrite a file like this:

161788  group_monitor.sgmops    4530    1293840320  1293840152
161789  group_atlas.atlas053    22350   1293840262  1293840152
161790  group_alice.alice017    210     1293840254  1293840159
161791  group_lhcb.pltlhc15     108277  1293949235  1293840159
161792  group_atlas.sgmatlas    35349   1293840251  1293840160

(where the last two fields are epoch time) ordered by the next to last field to this:

161792  group_atlas.sgmatlas    35349   1293840251  1293840160
161790  group_alice.alice017    210     1293840254  1293840159
161789  group_atlas.atlas053    22350   1293840262  1293840152
161788  group_monitor.sgmops    4530    1293840320  1293840152
161791  group_lhcb.pltlhc15     108277  1293949235  1293840159

As you can see, I used sorted(), which was introduced in v2.4, how can I rewrite the script for v2.3, so that it does that same thing. In addition, I want to convert the epoch time to the human-readable format, so the resultant file looks like this:

161792  group_atlas.sgmatlas    35349   01/01/11 00:04:11   01/01/11 00:02:40
161790  group_alice.alice017    210     01/01/11 00:04:14   01/01/11 00:02:39
161789  group_atlas.atlas053    22350   01/01/11 00:04:22   01/01/11 00:02:32

I know, this strftime("%d/%m/%y %H:%M:%S", gmtime()) can be used to convert the epoch time but I just can't figure out how can I apply that to the script to rewrite the file in that format.

Comments? Advice treasured!

In some cases, the epoch time comes as 3600, which is to indicate an unfinished business. I wanted to print aborted instead of 01/01/1970 for such a line. So, I changed the format_seconds_since_epoch() like this:

def format_seconds_since_epoch(t):
    if t == 3600:
        return "开发者_如何学Goaborted"
    else:
        return strftime("%d/%m/%y %H:%M:%S",datetime.fromtimestamp(t).timetuple())

which solved the problem. Is it the best that can be done in this regard? Cheers!!

file = open('readFile.txt', 'r')
lines = file.readlines()
file.close()
del file

lines = [line.split(' ') for line in lines]
lines.sort(lambda x,y: cmp(x[2], y[2])
lines = [' '.join(line) for line in lines]

In reply to your final query, you can create a datetime object from a time_t-like "seconds since the epoch" value using datetime.fromtimestamp, e.g.

from datetime import datetime
from time import strftime

def format_seconds_since_epoch(t):
    return strftime("%d/%m/%y %H:%M:%S",datetime.fromtimestamp(t).timetuple())

print format_seconds_since_epoch(1293840160)

So, putting that together with a slightly modified version of pynator's answer, you script might look like:

#!/usr/bin/env python

from datetime import datetime
from time import strftime
import os

def format_seconds_since_epoch(t):
    return strftime("%d/%m/%y %H:%M:%S",datetime.fromtimestamp(t).timetuple())

fin = open('readFile.txt', 'r')
lines = fin.readlines()
fin.close()
del fin

split_lines = [ line.split("\t") for line in lines ]

split_lines.sort( lambda a, b: cmp(int(a[-2]),int(b[-2])) )

fout = open('outfile.txt', 'w')
for split_line in split_lines:
    for i in (-2,-1):
        split_line[i] = format_seconds_since_epoch(int(split_line[i]))
    fout.write("\t".join(split_line)+os.linesep)

fout.close()
del fout

Note that using file as a variable name is a bad idea, since it shadows the built-in file type, so I changed them to fin and fout. (Even though you are deling the variables afterwards, it's still good style to avoid the name file, I think.)

In reply to your further question about the special "3600" value, your solution is fine. Personally, I would probably keep the format_seconds_since_epoch function as it is, so that it doesn't have a surprising special case and is more generally useful. You could create an additional wrapper function with the special case, or just change the split_line[i] = format_seconds_since_epoch(int(split_line[i])) line to:

entry = int(split_line[i])
if entry == 3600:
    split_line[i] = "aborted"
else:
    split_line[i] = format_seconds_since_epoch(entry)

... however I don't think there's much in the difference.

Incidentally, if this is more than a one-off task, I would see if you can use a later version of Python in the 2 series than 2.3, which is rather old now - they have lots of nice features that help one to write cleaner scripts.