BUG: enable resampling with NaT in PeriodIndex (GH 13224)

Author: winklerand

pandas/core/resample.py

@@ -1246,18 +1246,34 @@ def _get_period_bins(self, ax):
             raise TypeError('axis must be a PeriodIndex, but got '
                             'an instance of %r' % type(ax).__name__)
 
-        if not len(ax):
+        memb = ax.asfreq(self.freq, how=self.convention)
+        # NaT handling as in pandas._lib.lib.generate_bins_dt64()
+        nat_count = 0
+        if memb.hasnans:
+            import warnings
+            with warnings.catch_warnings():
+                warnings.filterwarnings('ignore', 'numpy equal will not check '
+                                                  'object identity')
+                nat_mask = memb.base == tslib.NaT
+                # raises "FutureWarning: numpy equal will not check object
+                # identity in the future. The comparison did not return the
+                # same result as suggested by the identity (`is`)) and will
+                # change."
+            nat_count = np.sum(nat_mask)
+            memb = memb[~nat_mask]
+
+        # if index contains no valid (non-NaT) values, return empty index
+        if not len(memb):
             binner = labels = PeriodIndex(
                 data=[], freq=self.freq, name=ax.name)
             return binner, [], labels
 
-        start = ax[0].asfreq(self.freq, how=self.convention)
-        end = ax[-1].asfreq(self.freq, how='end')
+        start = ax.min().asfreq(self.freq, how=self.convention)
+        end = ax.max().asfreq(self.freq, how='end')
 
         labels = binner = PeriodIndex(start=start, end=end,
                                       freq=self.freq, name=ax.name)
 
-        memb = ax.asfreq(self.freq, how=self.convention)
         i8 = memb.asi8
         freq_mult = self.freq.n
         # when upsampling to subperiods, we need to generate enough bins
@@ -1267,6 +1283,14 @@ def _get_period_bins(self, ax):
         rng += freq_mult
         bins = memb.searchsorted(rng, side='left')
 
+        if nat_count > 0:
+            # NaT handling as in pandas._lib.lib.generate_bins_dt64()
+            # shift bins by the number of NaT
+            bins += nat_count
+            bins = np.insert(bins, 0, nat_count)
+            binner = binner.insert(0, tslib.NaT)
+            labels = labels.insert(0, tslib.NaT)
+
         return binner, bins, labels
 
 

pandas/tests/test_resample.py

@@ -2870,6 +2870,45 @@ def test_upsampling_ohlc_freq_multiples(self):
         result = s.resample('12H', kind='period').ohlc()
         assert_frame_equal(result, expected)
 
+    def test_resample_with_nat(self):
+        # GH 13224
+        index = PeriodIndex([pd.NaT, '1970-01-01 00:00:00', pd.NaT,
+                             '1970-01-01 00:00:01', '1970-01-01 00:00:02'],
+                            freq='S')
+        frame = DataFrame([2, 3, 5, 7, 11], index=index)
+
+        index_1s = PeriodIndex(['1970-01-01 00:00:00', '1970-01-01 00:00:01',
+                                '1970-01-01 00:00:02'], freq='S')
+        frame_1s = DataFrame([3, 7, 11], index=index_1s)
+        result_1s = frame.resample('1s').mean()
+        assert_frame_equal(result_1s, frame_1s)
+
+        index_2s = PeriodIndex(['1970-01-01 00:00:00',
+                                '1970-01-01 00:00:02'], freq='2S')
+        frame_2s = DataFrame([5, 11], index=index_2s)
+        result_2s = frame.resample('2s').mean()
+        assert_frame_equal(result_2s, frame_2s)
+
+        index_3s = PeriodIndex(['1970-01-01 00:00:00'], freq='3S')
+        frame_3s = DataFrame([7], index=index_3s)
+        result_3s = frame.resample('3s').mean()
+        assert_frame_equal(result_3s, frame_3s)
+
+        pi = PeriodIndex(['1970-01-01 00:00:00', pd.NaT,
+                          '1970-01-01 00:00:02'], freq='S')
+        frame = DataFrame([2, 3, 5], index=pi)
+        expected_index = period_range(pi[0], periods=len(pi), freq=pi.freq)
+        expected = DataFrame([2, np.NaN, 5], index=expected_index)
+        result = frame.resample('1s').mean()
+        assert_frame_equal(result, expected)
+
+        pi = PeriodIndex([pd.NaT] * 3, freq='S')
+        frame = DataFrame([2, 3, 5], index=pi)
+        expected_index = PeriodIndex(data=[], freq=pi.freq)
+        expected = DataFrame([], index=expected_index)
+        result = frame.resample('1s').mean()
+        assert_frame_equal(result, expected)
+
 
 class TestTimedeltaIndex(Base, tm.TestCase):
     _index_factory = lambda x: timedelta_range