[Days 01–03] #100DaysOfCode

Dealing with Datetimes

Python datetimes

Let’s imagine we have a log file logfile looking like this:

INFO 2014-07-03T23:27:51 supybot Shutdown initiated.
ERROR 2014-07-03T23:24:31 supybot Invalid user dictionary file.
...
INFO 2014-07-03T23:31:22 supybot Shutdown initiated.

Part 1

Let’s propose a Python function that could parse a line of that file into a datetime object (think datetime(2014, 7, 3, 23, 27, 51) )

from datetime import datetime
def convert_to_datetime(line):
    date_line = datetime.fromisoformat(line.split(' ')[1])
    return date_line

Another way to achieve the same result (offered as a solution by the task designers):

def convert_to_datetime(line):
    timestamp = line.split()[1]
    date_str = '%Y-%m-%dT%H:%M:%S'
    return datetime.strptime(timestamp, date_str)

Now we are facing a question: what’s the difference between datetime.fromisoformat() and datetime.strptime(). Let’s see.

• datetime.fromisoformat(date_string) according to the Python docs returns a date object (year, month and day in an idealized calendar) corresponding to a date_string given in the format YYYY-MM-DD, so it recognizes this particular format only.
• datetime.strptime(date_string, format) returns a full datetime object (year, month, day, hour, minute, second, microsecond, tzinfo) corresponding to date_string, parsed according to format. You can see an example of such format syntax in the second function code above.

Part 2

Now let’s create a function that takes in the log file, identifies first and last system shutdown events and calculates a time difference between these two events.

First we process the logfile to obtain a list of its lines:

with open(logfile) as f:
    loglines = f.readlines()

Next, my suggestion of the required function:

def time_between_shutdowns(loglines):
    shutdown_list = [convert_to_datetime(el) for el in loglines if 'Shutdown initiated' in el]
    return shutdown_list[-1] - shutdown_list[0]

The function returns a timedelta object, reserved for storing duration, the difference between two dates or times. Below you’ll find a portion of the code that showcases timedelta capabilities.

from datetime import datetime
from datetime import timedelta
t = timedelta(days=4, hours=10)
t.days
#4
t.seconds
#36000
t.hours
#Traceback (most recent call last):
    #File "<pyshell#119>", line 1, in <module> t.hours
#AttributeError: 'datetime.timedelta' object has no attribute 'hours'
t.seconds / 60 / 60
#10.0
t.seconds / 3600
#10.0
#########
eta = timedelta(hours=6)
today = datetime.today()
today
#datetime.datetime(2018, 2, 19, 14, 55, 19, 197404)
today + eta
#datetime.datetime(2018, 2, 19, 20, 55, 19, 197404)
str(today + eta)
#'2018-02-19 20:55:19.197404'

You can find the practical task here