resample gives AmbiguousTimeError when index ends on a DST boundary

[JackKelly]

Here's the bug:

In [27]: idx = pd.date_range("2014-10-25 22:00:00", "2014-10-26 00:30:00", 
                             freq="30T", tz="Europe/London") 

In [28]: series = pd.Series(np.random.randn(len(idx)), index=idx)

In [31]: series
Out[31]: 
2014-10-25 22:00:00+01:00   -0.874014
2014-10-25 22:30:00+01:00    1.316258
2014-10-25 23:00:00+01:00   -1.334616
2014-10-25 23:30:00+01:00   -1.200390
2014-10-26 00:00:00+01:00   -0.341764
2014-10-26 00:30:00+01:00    1.509091
Freq: 30T, dtype: float64

In [29]: series.resample('30T')
---------------------------------------------------------------------------
AmbiguousTimeError                        Traceback (most recent call last)
<ipython-input-29-bb9e86068ce1> in <module>()
----> 1 series.resample('30T')

/usr/local/lib/python2.7/dist-packages/pandas/core/generic.pyc in resample(self, rule, how, axis, fill_method, closed, label, convention, kind, loffset, limit, base)
   3195                               fill_method=fill_method, convention=convention,
   3196                               limit=limit, base=base)
-> 3197         return sampler.resample(self).__finalize__(self)
   3198 
   3199     def first(self, offset):

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in resample(self, obj)
     83 
     84         if isinstance(ax, DatetimeIndex):
---> 85             rs = self._resample_timestamps()
     86         elif isinstance(ax, PeriodIndex):
     87             offset = to_offset(self.freq)

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in _resample_timestamps(self, kind)
    273         axlabels = self.ax
    274 
--> 275         self._get_binner_for_resample(kind=kind)
    276         grouper = self.grouper
    277         binner = self.binner

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in _get_binner_for_resample(self, kind)
    121             kind = self.kind
    122         if kind is None or kind == 'timestamp':
--> 123             self.binner, bins, binlabels = self._get_time_bins(ax)
    124         elif kind == 'timedelta':
    125             self.binner, bins, binlabels = self._get_time_delta_bins(ax)

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in _get_time_bins(self, ax)
    162         first, last = ax.min(), ax.max()
    163         first, last = _get_range_edges(first, last, self.freq, closed=self.closed,
--> 164                                        base=self.base)
    165         tz = ax.tz
    166         binner = labels = DatetimeIndex(freq=self.freq,

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in _get_range_edges(first, last, offset, closed, base)
    392         if (is_day and day_nanos % offset.nanos == 0) or not is_day:
    393             return _adjust_dates_anchored(first, last, offset,
--> 394                                           closed=closed, base=base)
    395 
    396     if not isinstance(offset, Tick):  # and first.time() != last.time():

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in _adjust_dates_anchored(first, last, offset, closed, base)
    459 
    460     return (Timestamp(fresult).tz_localize(first_tzinfo),
--> 461             Timestamp(lresult).tz_localize(first_tzinfo))
    462 
    463 

pandas/tslib.pyx in pandas.tslib.Timestamp.tz_localize (pandas/tslib.c:10535)()

pandas/tslib.pyx in pandas.tslib.tz_localize_to_utc (pandas/tslib.c:50297)()

AmbiguousTimeError: Cannot infer dst time from Timestamp('2014-10-26 01:00:00'), 
try using the 'ambiguous' argument

This is my hacky work-around:

In [35]: series.tz_convert('UTC').resample('30T').tz_convert('Europe/London')
Out[35]: 
2014-10-25 22:00:00+01:00   -0.874014
2014-10-25 22:30:00+01:00    1.316258
2014-10-25 23:00:00+01:00   -1.334616
2014-10-25 23:30:00+01:00   -1.200390
2014-10-26 00:00:00+01:00   -0.341764
2014-10-26 00:30:00+01:00    1.509091
Freq: 30T, dtype: float64

The bug disappears if the end date of the index is beyond the DST boundary:

In [37]: idx = pd.date_range("2014-10-25 22:00:00", "2014-10-26 02:30:00",
                             freq="30T", tz="Europe/London") 

In [38]: series = pd.Series(np.random.randn(len(idx)), index=idx)

In [39]: series.resample('30T')
Out[39]: 
2014-10-25 22:00:00+01:00   -0.626598
2014-10-25 22:30:00+01:00    1.799176
2014-10-25 23:00:00+01:00   -0.388075
2014-10-25 23:30:00+01:00    0.641487
2014-10-26 00:00:00+01:00   -0.488203
2014-10-26 00:30:00+01:00    0.477301
2014-10-26 01:00:00+01:00    0.040997
2014-10-26 01:30:00+01:00   -1.996542
2014-10-26 01:00:00+00:00   -0.016655
2014-10-26 01:30:00+00:00   -1.445823
2014-10-26 02:00:00+00:00   -0.713523
2014-10-26 02:30:00+00:00   -0.122274
Freq: 30T, dtype: float64

All is fine if the start date is on a DST boundary but the end date is beyond the boundary:

In [47]: idx = pd.date_range("2014-10-26 00:30:00", "2014-10-26 02:30:00",
                             freq="30T", tz="Europe/London") 

In [48]: series = pd.Series(np.random.randn(len(idx)), index=idx)

In [49]: series.resample('30T')
Out[49]: 
2014-10-26 00:30:00+01:00    2.371051
2014-10-26 01:00:00+01:00   -0.033473
2014-10-26 01:30:00+01:00   -0.988517
2014-10-26 01:00:00+00:00   -0.664475
2014-10-26 01:30:00+00:00    0.865772
2014-10-26 02:00:00+00:00   -1.051219
2014-10-26 02:30:00+00:00   -1.478477
Freq: 30T, dtype: float64

(As always, I must say a huge THANK YOU to everyone working on Pandas; it really is a great bit of software)

In [50]: pd.show_versions()

INSTALLED VERSIONS
------------------
commit: None
python: 2.7.9.final.0
python-bits: 64
OS: Linux
OS-release: 3.19.0-16-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_GB.UTF-8

pandas: 0.16.1
nose: 1.3.4
Cython: 0.22
numpy: 1.9.2
scipy: 0.14.0
statsmodels: 0.6.1
IPython: 3.1.0
sphinx: 1.2.3
patsy: 0.3.0
dateutil: 2.4.2
pytz: 2015.2
bottleneck: None
tables: 3.1.1
numexpr: 2.4
matplotlib: 1.4.3
openpyxl: None
xlrd: 0.9.2
xlwt: None
xlsxwriter: None
lxml: None
bs4: 4.3.2
html5lib: 0.999
httplib2: 0.9
apiclient: None
sqlalchemy: None
pymysql: None
psycopg2: 2.5.3 (dt dec pq3 ext)

Possibly related:

• Fixes to resample over DST boundaries #9623
• BUG: Resampling a TZ-aware Series fails if rule='W': "Values falls before first bin" #9119
• AmbiguousTimeError exception when resampling near DST change date #8744
• BUG/API: master issue for resample on DST days #5172

[JackKelly]

I saw that this bug is still open so I thought I should re-test Pandas 0.16.2 to see if the bug had been fixed along the way. Unfortunately this bug still exists. I have included the new traceback below (because it has different line numbers to the traceback above):

In [79]: import pandas as pd

In [80]: idx = pd.date_range("2014-10-25 22:00:00", "2014-10-26 00:30:00",freq="30T", 
          tz="Europe/London") 

In [81]: series = pd.Series(np.random.randn(len(idx)), index=idx)

In [82]: series
Out[82]: 
2014-10-25 22:00:00+01:00   -1.315553
2014-10-25 22:30:00+01:00    0.294073
2014-10-25 23:00:00+01:00    0.067067
2014-10-25 23:30:00+01:00    0.710251
2014-10-26 00:00:00+01:00   -0.192490
2014-10-26 00:30:00+01:00    0.661763
Freq: 30T, dtype: float64

In [83]: series.resample('30T')
---------------------------------------------------------------------------
AmbiguousTimeError                        Traceback (most recent call last)
<ipython-input-83-bb9e86068ce1> in <module>()
----> 1 series.resample('30T')

/usr/local/lib/python2.7/dist-packages/pandas/core/generic.pyc in resample(self, rule, how, axis, fill_method, closed, label, convention, kind, loffset, limit, base)
   3264                               fill_method=fill_method, convention=convention,
   3265                               limit=limit, base=base)
-> 3266         return sampler.resample(self).__finalize__(self)
   3267 
   3268     def first(self, offset):

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in resample(self, obj)
     80 
     81         if isinstance(ax, DatetimeIndex):
---> 82             rs = self._resample_timestamps()
     83         elif isinstance(ax, PeriodIndex):
     84             offset = to_offset(self.freq)

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in _resample_timestamps(self, kind)
    270         axlabels = self.ax
    271 
--> 272         self._get_binner_for_resample(kind=kind)
    273         grouper = self.grouper
    274         binner = self.binner

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in _get_binner_for_resample(self, kind)
    118             kind = self.kind
    119         if kind is None or kind == 'timestamp':
--> 120             self.binner, bins, binlabels = self._get_time_bins(ax)
    121         elif kind == 'timedelta':
    122             self.binner, bins, binlabels = self._get_time_delta_bins(ax)

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in _get_time_bins(self, ax)
    159         first, last = ax.min(), ax.max()
    160         first, last = _get_range_edges(first, last, self.freq, closed=self.closed,
--> 161                                        base=self.base)
    162         tz = ax.tz
    163         binner = labels = DatetimeIndex(freq=self.freq,

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in _get_range_edges(first, last, offset, closed, base)
    389         if (is_day and day_nanos % offset.nanos == 0) or not is_day:
    390             return _adjust_dates_anchored(first, last, offset,
--> 391                                           closed=closed, base=base)
    392 
    393     if not isinstance(offset, Tick):  # and first.time() != last.time():

/usr/local/lib/python2.7/dist-packages/pandas/tseries/resample.pyc in _adjust_dates_anchored(first, last, offset, closed, base)
    456 
    457     return (Timestamp(fresult).tz_localize(first_tzinfo),
--> 458             Timestamp(lresult).tz_localize(first_tzinfo))
    459 
    460 

pandas/tslib.pyx in pandas.tslib.Timestamp.tz_localize (pandas/tslib.c:10551)()

pandas/tslib.pyx in pandas.tslib.tz_localize_to_utc (pandas/tslib.c:50619)()

AmbiguousTimeError: Cannot infer dst time from Timestamp('2014-10-26 01:00:00'), 
try using the 'ambiguous' argument

[jreback]

if it were fixed it would be closed

pull requests are welcome

[jreback]

cc @rockg

[linar-jether]

This happens also when index starts on DST boundary,

df = pd.DataFrame(columns=['a', 'b', 'c'], index=pd.date_range('2014-03-09 03:00', '2015-03-09 03:00', freq='H', tz='America/Chicago')).assign(a=np.random.rand(), b=np.random.rand(), c=np.random.rand())
df.resample('H', label='right', closed='right').sum()

results in: NonExistentTimeError: 2014-03-09 02:00:00