tags:

views:

140

answers:

6

In Python how do I find all the missing days in a sorted list of dates?

+2  A: 

Sort the list of dates and iterate over it, remembering the previous entry. If the difference between the previous and current entry is more than one day, you have missing days.

Here's one way to implement it:

from datetime import date, timedelta
from itertools import tee, izip

def pairwise(iterable):
    "s -> (s0,s1), (s1,s2), (s2, s3), ..."
    a, b = tee(iterable)
    b.next()
    return izip(a, b)

def missing_dates(dates):
    for prev, curr in pairwise(sorted(dates)):
        i = prev
        while i + timedelta(1) < curr:
            i += timedelta(1)
            yield i

dates = [ date(2010, 1, 8),
          date(2010, 1, 2),
          date(2010, 1, 5),
          date(2010, 1, 1),
          date(2010, 1, 7) ]

for missing in missing_dates(dates):
    print missing

Output:

2010-01-03
2010-01-04
2010-01-06

Performance is O(n*log(n)) where n is the number of days in the span when the input is unsorted. As your list is already sorted, it will run in O(n).

Mark Byers
The list of dates is already sorted
gnibbler
@gnibbler: That's great... then this algorithm runs in O(n) time.
Mark Byers
Mine is O(n) too. Had I not used date_set it would have had worst case performance of O(n*n) though :)
gnibbler
+1  A: 

Put the dates in a set and then iterate from the first date to the last using datetime.timedelta(), checking for containment in the set each time.

Ignacio Vazquez-Abrams
+2  A: 
>>> from datetime import datetime, timedelta
>>> date_list = [datetime(2010, 2, 23),datetime(2010, 2, 24),datetime(2010, 2, 25),datetime(2010, 2, 26),datetime(2010, 3, 1),datetime(2010, 3, 2)]
>>> 
>>> date_set=set(date_list)         # for faster membership tests than list
>>> one_day = timedelta(days=1)
>>> 
>>> test_date = date_list[0]
>>> missing_dates=[]
>>> while test_date < date_list[-1]:
...     if test_date not in date_set:
...         missing_dates.append(test_date)
...     test_date += one_day
... 
>>> print missing_dates
[datetime.datetime(2010, 2, 27, 0, 0), datetime.datetime(2010, 2, 28, 0, 0)]

This also works for datetime.date objects, but the OP says the list is datetime.datetime objects

gnibbler
+1: Seems good. Nice touch with using the parameter name on the timedelta(days=1) for clarity.
Mark Byers
A: 
import datetime

DAY = datetime.timedelta(days=1)
# missing dates: a list of [start_date, end)
missing = [(d1+DAY, d2) for d1, d2 in zip(dates, dates[1:]) if (d2 - d1) > DAY]

def date_range(start_date, end, step=DAY):
    d = start_date
    while d < end:
        yield d
        d += step

missing_dates = [d for d1, d2 in missing for d in date_range(d1, d2)]
J.F. Sebastian
A: 

Using a list comprehension

>>> from datetime import date, timedelta
>>> d = [date(2010, 2, 23),date(2010, 2, 24),date(2010, 2, 25),date(2010, 2, 26),date(2010, 3, 1),date(2010, 3, 2)]
>>> date_set=set(d)
>>> missing = [x for x in (d[0]+timedelta(x) for x in range((d[-1]-d[0]).days)) if x not in date_set]

>>> missing
[datetime.date(2010, 2, 27), datetime.date(2010, 2, 28)]
gnibbler
+4  A: 

using sets

>>> from datetime import date, timedelta
>>> d = [date(2010, 2, 23),date(2010, 2, 24),date(2010, 2, 25),
         date(2010, 2, 26),date(2010, 3, 1),date(2010, 3, 2)]
>>> date_set = set(d[0]+timedelta(x) for x in range((d[-1]-d[0]).days))
>>> missing = sorted(date_set-set(d))
>>> missing
[datetime.date(2010, 2, 27), datetime.date(2010, 2, 28)]
>>> 
gnibbler
Appreciate your responses, I like this one :) Thanks.
Vishal