ENH: Introduces keyword in resample to handle open intervals. pandas-dev#10449

Author: dagru

pandas/core/generic.py

@@ -3557,7 +3557,7 @@ def between_time(self, start_time, end_time, include_start=True,
 
     def resample(self, rule, how=None, axis=0, fill_method=None, closed=None,
                  label=None, convention='start', kind=None, loffset=None,
-                 limit=None, base=0):
+                 limit=None, base=0, was_closed=None):
         """
         Convenience method for frequency conversion and resampling of regular
         time-series data.
@@ -3578,6 +3578,11 @@ def resample(self, rule, how=None, axis=0, fill_method=None, closed=None,
             For frequencies that evenly subdivide 1 day, the "origin" of the
             aggregated intervals. For example, for '5min' frequency, base could
             range from 0 through 4. Defaults to 0
+        was_closed: boolean
+            Interprets the given time-series as sampling points of a left-closed right-open interval (or vice 
+            versa, if closed=='right'). Resamples up to (but excluding) the assumed end-point of the open side of the interval.
+            
+            .. versionadded:: 0.18.0
 
 
         Examples
@@ -3676,15 +3681,44 @@ def resample(self, rule, how=None, axis=0, fill_method=None, closed=None,
         2000-01-01 00:03:00    17
         2000-01-01 00:06:00    26
         Freq: 3T, dtype: int64
-
+        
+        Use was_closed to resample up to (excluding) the next sampling point
+        
+        >>> series.resample("15s").pad()[24:]
+        2000-01-01 00:06:00    6
+        2000-01-01 00:06:15    6
+        2000-01-01 00:06:30    6
+        2000-01-01 00:06:45    6
+        2000-01-01 00:07:00    7
+        2000-01-01 00:07:15    7
+        2000-01-01 00:07:30    7
+        2000-01-01 00:07:45    7
+        2000-01-01 00:08:00    8
+        Freq: 15S, dtype: int64
+        
+        >>> series.resample("15s", was_closed=True).pad()[24:]
+        2000-01-01 00:06:00    6
+        2000-01-01 00:06:15    6
+        2000-01-01 00:06:30    6
+        2000-01-01 00:06:45    6
+        2000-01-01 00:07:00    7
+        2000-01-01 00:07:15    7
+        2000-01-01 00:07:30    7
+        2000-01-01 00:07:45    7
+        2000-01-01 00:08:00    8
+        2000-01-01 00:08:15    8
+        2000-01-01 00:08:30    8
+        2000-01-01 00:08:45    8
+        Freq: 15S, dtype: int64
+        
         """
         from pandas.tseries.resample import resample
 
         axis = self._get_axis_number(axis)
         r = resample(self, freq=rule, label=label, closed=closed,
                      axis=axis, kind=kind, loffset=loffset,
                      fill_method=fill_method, convention=convention,
-                     limit=limit, base=base)
+                     limit=limit, base=base, was_closed=was_closed)
 
         # deprecation warnings
         # but call methods anyhow

pandas/tseries/resample.py

@@ -48,7 +48,7 @@ class Resampler(_GroupBy):
 
     # to the groupby descriptor
     _attributes = ['freq', 'axis', 'closed', 'label', 'convention',
-                   'loffset', 'base', 'kind']
+                   'loffset', 'base', 'kind', 'was_closed']
 
     # API compat of allowed attributes
     _deprecated_valids = _attributes + ['_ipython_display_', '__doc__',
@@ -721,7 +721,8 @@ class TimeGrouper(Grouper):
     def __init__(self, freq='Min', closed=None, label=None, how='mean',
                  nperiods=None, axis=0,
                  fill_method=None, limit=None, loffset=None, kind=None,
-                 convention=None, base=0, **kwargs):
+                 convention=None, base=0,
+                 was_closed=None, **kwargs):
         freq = to_offset(freq)
 
         end_types = set(['M', 'A', 'Q', 'BM', 'BA', 'BQ', 'W'])
@@ -752,6 +753,8 @@ def __init__(self, freq='Min', closed=None, label=None, how='mean',
         self.limit = limit
         self.base = base
 
+        self.was_closed = was_closed
+
         # always sort time groupers
         kwargs['sort'] = True
 
@@ -852,7 +855,8 @@ def _get_time_bins(self, ax):
         first, last = ax.min(), ax.max()
         first, last = _get_range_edges(first, last, self.freq,
                                        closed=self.closed,
-                                       base=self.base)
+                                       base=self.base,
+                                       was_closed=self.was_closed, was_freq=ax.freq)
         tz = ax.tz
         binner = labels = DatetimeIndex(freq=self.freq,
                                         start=first.replace(tzinfo=None),
@@ -978,7 +982,32 @@ def _take_new_index(obj, indexer, new_index, axis=0):
         raise ValueError("'obj' should be either a Series or a DataFrame")
 
 
-def _get_range_edges(first, last, offset, closed='left', base=0):
+def _get_range_edges(first, last, offset, closed='left', base=0,
+                     was_closed=False, was_freq=None):
+
+    if was_closed is True:
+      if not was_freq:
+        raise ValueError("was_closed requires equidstant index.")
+
+      virt_first = first - (closed == 'right') * was_freq
+      virt_last = last + (closed == 'left') * was_freq
+
+      diff_nano = (virt_last - virt_first).value
+
+      if closed == "left":
+          tmp_first = virt_first
+          tmp_last = virt_last
+          if diff_nano % offset.nanos > 0:
+              tmp_last = tmp_last + offset
+      else:
+          periods = (virt_last.value - virt_first.value) // offset.nanos
+          if diff_nano % offset.nanos > 0:
+              periods += 1
+          tmp_first = virt_last - periods * offset
+          tmp_last = virt_last
+
+      return tmp_first, tmp_last
+
     if isinstance(offset, compat.string_types):
         offset = to_offset(offset)
 
@@ -1000,7 +1029,6 @@ def _get_range_edges(first, last, offset, closed='left', base=0):
         first = Timestamp(offset.rollback(first))
     else:
         first = Timestamp(first - offset)
-
     last = Timestamp(last + offset)
 
     return first, last

pandas/tseries/tests/test_resample.py

@@ -1548,6 +1548,24 @@ def test_resmaple_dst_anchor(self):
                                        freq='D', tz='Europe/Paris')),
             'D Frequency')
 
+    def test_wasclosed(self):
+        df = pd.Series(1, index=pd.date_range(
+            "2015-01-01 00:00:00", "2015-01-01 02:00:00", freq="1h"))
+        tests = [17, 30, 55, 60, 77, 120, 127, 243]
+        expected_periods = [(60 % i > 0) + 180 // i for i in tests]
+        expected_freqs = ["%imin" % i for i in tests]
+
+        for i in range(len(tests)):
+            goal_left = pd.date_range(
+                start="2015-01-01 00:00:00", freq=expected_freqs[i], periods=expected_periods[i])
+            goal_right = pd.DatetimeIndex(
+                end="2015-01-01 02:00:00", freq=expected_freqs[i], periods=expected_periods[i])
+            resampled_left = df.resample(
+                expected_freqs[i], closed="left", was_closed=True).ffill().index
+            resampled_right = df.resample(
+                expected_freqs[i], closed="right", was_closed=True).ffill().index
+            tm.assert_index_equal(resampled_left, goal_left)
+            tm.assert_index_equal(resampled_right, goal_right)
 
 def _simple_ts(start, end, freq='D'):
     rng = date_range(start, end, freq=freq)