Building a list of dates between two dates

My solution to this feels 'icky' and I've got calendar math falling out of my ears after working on similar problems for a week so I can't think straight about this.

Is there a better way to code this?

```
import datetime
from dateutil.relativedelta import relativedelta

def date_count(start, end, day_of_month=1):
"""
Return a list of datetime.date objects that lie in-between start and end.

The first element of the returned list will always be start and the last
element in the returned list will always be:
datetime.date(end.year, end.month, day_of_month)

If start.day is equal to day_of_month the second element will be:
start + 1 month

If start.day is after day_of_month then the second element will be:
the day_of_month in the next month

If start.day is before day_of_month then the second element will be:
datetime.date(start.year, start.month, day_of_month)

>>> start = datetime.date(2012, 1, 15)
>>> end = datetime.date(2012, 4, 1)
>>> date_count(start, end, day_of_month=1) #doctest: +NORMALIZE_WHITESPACE
[datetime.date(2012, 1, 15), datetime.date(2012, 2, 1),
datetime.date(2012, 3, 1), datetime.date(2012, 4, 1)]

Notice that it's not a full month between the first two elements in the
list.

If you have a start day before day_of_month:
>>> start = datetime.date(2012, 1, 10)
>>> end = datetime.date(2012, 4, 1)
>>> date_count(start, end, day_of_month=15) #doctest: +NORMALIZE_WHITESPACE
[datetime.date(2012, 1, 10), datetime.date(2012, 1, 15),
datetime.date(2012, 2, 15), datetime.date(2012, 3, 15),
datetime.date(2012, 4, 15)]

Notice that it's not a full month between the first two elements in the
list and that the last day is rounded to
datetime.date(end.year, end.month, day_of_month)
"""
last_element = datetime.date(end.year, end.month, day_of_month)

if start.day == day_of_month:
second_element = start + relativedelta

How about...

def date_count(start, end, day_of_month=1):
    dates = [start]
    next_date = start.replace(day=day_of_month)
    if day_of_month > start.day:
        dates.append(next_date)
    while next_date < end.replace(day=day_of_month):
        next_date += relativedelta(next_date, months=+1)
        dates.append(next_date)
    return dates

And by the way it seems like a nice opportunity to use yield, if you wanted to.

def date_count2(start, end, day_of_month=1):
    yield start
    next_date = start.replace(day=day_of_month)
    if day_of_month > start.day:
        yield next_date
    while next_date < end.replace(day=day_of_month):
        next_date += relativedelta(next_date, months=+1)
        yield next_date

Another possibility - discard the first value if it is earlier than the start date:

def date_count(start, end, day_of_month=1):
    dates = [start.replace(day=day_of_month)]
    while dates[-1]  start:
        return [start] + dates
    else:
        return [start] + dates[1:]

Or use a list comprehension to iterate over the number of months between start and end.

def date_count(start, end, day_of_month=1):
    round_start = start.replace(day=day_of_month)
    gap = end.year * 12 + end.month - start.year * 12 - start.month + 1
    return [start] + [round_start + relativedelta(round_start, months=i) 
                      for i in range(day_of_month <= start.day, gap)]

Finally, I don't know dateutil but it seems you can use rrule:

from dateutil import rrule
def date_count(start, end, day_of_month=1):
    yield start
    for date in rrule(MONTHLY, 
                      dtstart=start.replace(day=day_of_month),
                      until=end.replace(day=day_of_month)):
        if date > start:
            yield date

StackExchange Code Review Q#19903, answer score: 3