Bug: CustomBusinessHour bug with holidays

[ustcscgyer]

Code Sample, a copy-pastable example if possible

# Your code here
import pandas as pd # My pandas version is 0.19.2

from pandas.tseries.holiday import USFederalHolidayCalendar
from pandas.tseries.offsets import CustomBusinessHour
import datetime as dt
bhour_us = CustomBusinessHour(calendar=USFederalHolidayCalendar())

# Friday before MLK Day
t0 = dt.datetime(2014, 1, 17, 15)

# This is still right
# Expected Result: Timestamp('2014-01-21 14:00:00')
# print(t0 + bhour_us * 7 )

# This is wrong
# Expected Result: Timestamp('2014-01-21 15:00:00')
print(t0 + bhour_us * 8)
# But the code generates: Timestamp('2014-01-20 15:00:00')

Problem description

I used a Timestamp one day prior to a Holiday as t0. When offset it by number of hours that is less than the number of hours of a business day, pandas correctly accounts for the holidays, but when the number of hours in the offset is equal or larger than number of hours of a business day, pandas gives one day prior to the correct result.

Expected Output

Timestamp('2014-01-21 15:00:00')

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.4.5.final.0
python-bits: 64
OS: Windows
OS-release: 10
machine: AMD64
processor: Intel64 Family 6 Model 78 Stepping 3, GenuineIntel
byteorder: little
LC_ALL: None
LANG: None
LOCALE: None.None

pandas: 0.19.2
nose: 1.3.7
pip: 9.0.1
setuptools: 27.2.0
Cython: 0.25.2
numpy: 1.11.3
scipy: 0.18.1
statsmodels: None
xarray: None
IPython: 5.1.0
sphinx: 1.5.1
patsy: 0.4.1
dateutil: 2.6.0
pytz: 2016.10
blosc: None
bottleneck: 1.2.0
tables: 3.2.2
numexpr: 2.6.1
matplotlib: 2.0.0
openpyxl: 2.4.1
xlrd: 1.0.0
xlwt: 1.2.0
xlsxwriter: 0.9.6
lxml: 3.7.2
bs4: 4.5.3
html5lib: None
httplib2: None
apiclient: None
sqlalchemy: 1.1.5
pymysql: None
psycopg2: None
jinja2: 2.9.4
boto: 2.45.0
pandas_datareader: None

[TomAugspurger]

Yeah, these two look very strange:

In [34]: t0 + bhour_us * 7
Out[34]: Timestamp('2014-01-21 14:00:00')

In [35]: t0 + bhour_us * 8
Out[35]: Timestamp('2014-01-20 15:00:00')

Would welcome a PR if you have time to dig into what's going on.

[alanbato]

Claiming this bug :)

[alanbato]

So far I could identify the offending part of the code in offsets.py:837

if bd != 0:
    skip_bd = BusinessDay(n=bd)
    # midnight business hour may not on BusinessDay
    if not self.next_bday.onOffset(other):
        remain = other - self._prev_opening_time(other)
        other = self._next_opening_time(other + skip_bd) + remain
    else:
        other = other + skip_bd

The call to self.next_bday.onOffset(other) returns True, so the code else clause returns other + skip_bd as the result of the offsetting. However, if the code in the if clause executed, it would return the correct result.

If anyone has any ideas on how to refactor/fix this part, I'd be happy to hear them.

[mroeschke]

This looks to be fixed on master. Could use a test

In [4]: import pandas as pd # My pandas version is 0.19.2
   ...:
   ...: from pandas.tseries.holiday import USFederalHolidayCalendar
   ...: from pandas.tseries.offsets import CustomBusinessHour
   ...: import datetime as dt
   ...: bhour_us = CustomBusinessHour(calendar=USFederalHolidayCalendar())
   ...:
   ...: # Friday before MLK Day
   ...: t0 = dt.datetime(2014, 1, 17, 15)

In [5]: print(t0 + bhour_us * 8)
2014-01-21 15:00:00

[tqa236]

I can help to add the tests. Can you tell me which file I should add them to?