Python-按正则表达式日期匹配对文件中的行进行排序

Question

I have a file (based on a class project) of scraped Tweets. At this point lines in the file look like:

@soandso something something <a href="http://pic.twitter.com/aphoto</a><a href="a link" target="_blank">Permalink</a> 1:40 PM - 17 Feb 2016<br><br>
@soandso something something <a href="http://pic.twitter.com/aphoto</a><a href="a link" target="_blank">Permalink</a> 1:32 PM - 16 Feb 2016<br><br>

I'm trying to sort the lines in the file by date. This is what I've cobbled together so far.

import re
from datetime import datetime

when = re.compile(r".+</a>(.+)<br><br>")

with open('tweets.txt','r+') as outfile:
    sortme = outfile.read()

    for match in re.finditer(when, sortme):
        tweet = match.group(0)
        when = match.group(1)
        when = datetime.strptime(when, " %I:%M %p - %d %b %Y")
        print when

Which will print out all the dates in the lines having converted the format from 1:40 PM - 17 Feb 2016 to 2016-02-17 13:40:00, which I believe is a datetime. I have searched high and low over the last few days for clues about how I'd then sort all the lines in the file by datetime. Thanks for your help!

Answer 1

I have searched high and low over the last few days for clues about how I'd then sort all the lines in the file by datetime.

def get_time(line):
    match = re.search(r"</a>\s*(.+?)\s*<br><br>", line)
    if match:
        return datetime.strptime(match.group(1), "%I:%M %p - %d %b %Y")
    return datetime.min

lines.sort(key=get_time)

It assumes that the time is monotonous in the given time period (eg, no DST transitions) otherwise you should convert the input time to UTC (or POSIX timestamp) first.

Answer 2

It seems you have already solved the regex problem... so to convert your datetime into a measurable quantity convert to seconds like so:

import time
time.mktime(when.timetuple())

then for sorting you can make a lot off different routes. the simplest example is:

import operator
s = [("ab",50),("cd",100),("ef",15)]
print sorted(s,key=operator.itemgetter(1))
## [('ef', 15), ('ab', 50), ('cd', 100)]