dealing with datetime while parsing a csv

[erbrose]

Hello all!
I am parsing a csv file and one of the fields is a date time field that looks something like this
2010-01-15 23:15:30
year-month-day hour24:minute:second

as i loop through this csv i am going to need to do some time arithmetic. my question is, how do i turn that field from a string in my list to a datetime object?

from datatime import datetime
TmpArr = []
reader = open("c:/test.csv",'r')
for line in reader:
    TmpArr = line.split(',')
    #now TmpArr[1] contains the datatime data
    #i've tried this
    dt = datetime(TmpArr[1])
    #but get a needs an integer error
    #i've tried this too
    dt = datetime.strtime(TmpArr[1],"%Y-%M-%D %h:%m:%s")
    # then i get a needs datetime object
 

lost and would greatly appreciate anyones help with this... Using Python 2.6.3
Thanks ahead of time.
Cheers,
Eric

[bvdet]

You have to go to the time module to get what you need.

import datetime, time
 
s = "2010-01-15 23:15:30"
print repr(datetime.datetime(*time.strptime(s, "%Y-%m-%d %H:%M:%S")[:6]))

Output:
>>> datetime.datetime(2010, 1, 15, 23, 15, 30)

[erbrose]

Thanks so much...
now seeing as i don't really see something like this in the documentation, could you explain this line a bit more?
datetime.datetime(*time.strptime(s, "%Y-%m-%d %H:%M:%S")[:6])

I understand calling datetime.datetime, but why the * before time? and also what exactly is the [:6] pertaining too
Thanks again for your help!!!

[bvdet]

time.strptime returns a struct_time object with more attributes than datetime.datetime can handle, therefore the slice ([:6]). The asterisk expands the remaining tuple into the individual tuple elements. Example:

>>> time.strptime(s, "%Y-%m-%d %H:%M:%S")
(2010, 1, 15, 23, 15, 30, 4, 15, -1)
>>> time.strptime(s, "%Y-%m-%d %H:%M:%S")[:6]
(2010, 1, 15, 23, 15, 30)
>>> datetime.datetime(time.strptime(s, "%Y-%m-%d %H:%M:%S")[:6])
Traceback (most recent call last):
  File "<interactive input>", line 1, in ?
TypeError: function takes at least 3 arguments (1 given)
>>> datetime.datetime(*time.strptime(s, "%Y-%m-%d %H:%M:%S")[:6])
datetime.datetime(2010, 1, 15, 23, 15, 30)
>>> 

[erbrose]

Thanks, that does make more sense now!