This is an unambigous win over the current code.

To see what the best thing to do is, I made two versions, parserinfo does the min call in the constructor and stores it as a private attribute, and parserinfo2 is the code as presented here. and parserinfo3 is the code as is now:

>>> i = parserinfo()
>>> i2 = parserinfo2()
>>> i3 = parserinfo3()

>>> %timeit i.weekday('Wed')
1000000 loops, best of 3: 627 ns per loop

>>> %timeit i2.weekday('Wed')
1000000 loops, best of 3: 393 ns per loop

>>> %timeit i3.weekday('Wed')
100000 loops, best of 3: 3.4 µs per loop

>>> %timeit i.weekday('A')
1000000 loops, best of 3: 285 ns per loop

>>> %timeit i2.weekday('A')
1000000 loops, best of 3: 724 ns per loop

>>> %timeit i3.weekday('A')
100000 loops, best of 3: 2.83 µs per loop

When the string passed to weekday has length >= the shortest string in in the weekday keys, this version is best. When it's <= the min, the cached version is best. The version we have now performs worst in all cases.

I'm not sure all the code paths for the common case, but I'm guessing that it's actually kinda rare to hit the case where weekday or month are passed 2-character strings, so this approach seems the best one to me.

Yah, once we get to the point of trying to out-perform a dict lookup, it's time to go read a book.

I suppose we could have done self._weekdays.get(name.lower(), None) for weekdays.