I'm trying to calculate the Julian Day, given a UTC year, month and date in the Gregorian calendar. I tried using the formula on Wikipedia, but that doesn't work. Consider 2010-01-31 and 2010-02-01. These dates are exactly one day apart, but their JDNs, according to the formula on Wikipedia, are 2455230 and 2455229, respectively.
In fact, it seems that this issue appears at the beginning and end of February of every year. I wrote a Python script to go through every day from 2010-2020, and these are all similar discrepancies:
ERROR: 2010-01-31 00:00:00 (2455230) vs 2010-02-01 00:00:00 (2455229)
ERROR: 2010-02-28 00:00:00 (2455256) vs 2010-03-01 00:00:00 (2455259)
ERROR: 2011-01-31 00:00:00 (2455595) vs 2011-02-01 00:00:00 (2455594)
ERROR: 2011-02-28 00:00:00 (2455621) vs 2011-03-01 00:00:00 (2455624)
ERROR: 2012-01-31 00:00:00 (2455960) vs 2012-02-01 00:00:00 (2455959)
ERROR: 2012-02-29 00:00:00 (2455987) vs 2012-03-01 00:00:00 (2455989)
ERROR: 2013-01-31 00:00:00 (2456325) vs 2013-02-01 00:00:00 (2456325)
ERROR: 2013-02-28 00:00:00 (2456352) vs 2013-03-01 00:00:00 (2456355)
ERROR: 2014-01-31 00:00:00 (2456691) vs 2014-02-01 00:00:00 (2456690)
ERROR: 2014-02-28 00:00:00 (2456717) vs 2014-03-01 00:00:00 (2456720)
ERROR: 2015-01-31 00:00:00 (2457056) vs 2015-02-01 00:00:00 (2457055)
ERROR: 2015-02-28 00:00:00 (2457082) vs 2015-03-01 00:00:00 (2457085)
ERROR: 2016-01-31 00:00:00 (2457421) vs 2016-02-01 00:00:00 (2457420)
ERROR: 2016-02-29 00:00:00 (2457448) vs 2016-03-01 00:00:00 (2457450)
ERROR: 2017-01-31 00:00:00 (2457786) vs 2017-02-01 00:00:00 (2457786)
ERROR: 2017-02-28 00:00:00 (2457813) vs 2017-03-01 00:00:00 (2457816)
ERROR: 2018-01-31 00:00:00 (2458152) vs 2018-02-01 00:00:00 (2458151)
ERROR: 2018-02-28 00:00:00 (2458178) vs 2018-03-01 00:00:00 (2458181)
ERROR: 2019-01-31 00:00:00 (2458517) vs 2019-02-01 00:00:00 (2458516)
ERROR: 2019-02-28 00:00:00 (2458543) vs 2019-03-01 00:00:00 (2458546)
The JDNs also do not match those on NASA's calculator, as it gives 2010-01-31 to be JD 2455228.
Since Wikipedia isn't always the most reliable resource, I scoured the internet for another website with a formula, and all I found was this page, which gives me the same issue, with a slightly different, step-by-step formula.
Am I doing something wrong? I'm sorry if I am posting this in the wrong community, if there is a better place please let me know.
The code I used to get the list above is below, with the Wikipedia algorithm:
import datetime
prev_d = None
prev = 0
def calc(_d):
global prev_d
global prev
y = _d.year
m = _d.month
d = _d.day
# Formula from Wikipedia
jd = ((1461 * (y + 4800 + (m - 14) // 12)) // 4 + (367 * (m - 2 - 12 * ((m - 14) // 12))) // 12 - (3 * ((y + 4900 + (m - 14) // 12) // 100)) // 4 + d - 32075)
if jd - prev != 1:
print("ERROR: {} ({}) vs {} ({})".format(prev_d, prev, _d, jd))
prev = jd
prev_d = _d
# Iterate day-by-day, starting 2010-01-01 UTC
for d in range(1262304000, 1262304000 + 60 * 60 * 24 * 365 * 10, 60 * 60 * 24):
calc(datetime.datetime.utcfromtimestamp(d))
```