Update:
I am sorry for my careless, mixing up the word parse and parser. This question should be deleted. But since someone answered it and received reputations, I kept it here. Sorry again.
What are the differences between (import dateutil.parser)
>>> import dateutil.parser
>>> t = dateutil.parser.parser("2012-01-19 17:21:00 BRST")
>>> type(t)
<class 'dateutil.parser.parser'>
and (from dateutil.parser import parse)
>>> from dateutil.parser import parse
>>> t = parse("2012-01-19 17:21:00 BRST")
>>> type(t)
<type 'datetime.datetime'>
Can anyone explain the differences between import dateutil.parser and from dateutil.parser import parse?
The problem is that you are actually calling the constructor for the parser object, not the parse method. You can either call dateutil.parser.parse or instantiate a dateutil.parser.parser object and call its parse() method.
>>> import dateutil.parser
>>> t = dateutil.parser.parse("2012-01-19 17:21:00 BRST")
>>> type(t)
datetime.datetime
>>> t
datetime.datetime(2012, 1, 19, 17, 21)
Generally you can construct a parser object with a dateutil.parser.parserinfo object, but since you're not actually using the parser object, it's not throwing an error when it detects that you've passed it a string instead.