.resample now supports loffset. (#2608)

* .resample now supports loffset.

* Update whats-new.rst

* Fix for pandas 0.19.2

* doc update.

* Review comments.

Author: dcherian

doc/whats-new.rst

@@ -45,6 +45,9 @@ Enhancements
   "dayofyear" and "dayofweek" accessors (:issue:`2597`).  By `Spencer Clark
   <https://github.com/spencerkclark>`_.
 - Support Dask ``HighLevelGraphs`` by `Matthew Rocklin <https://matthewrocklin.com>`_.
+- :py:meth:`DataArray.resample` and :py:meth:`Dataset.resample` now supports the
+  ``loffset`` kwarg just like Pandas.
+  By `Deepak Cherian <https://github.com/dcherian>`_
 
 
 Bug fixes

xarray/core/common.py

@@ -592,7 +592,7 @@ def rolling(self, dim=None, min_periods=None, center=False, **dim_kwargs):
                                  center=center)
 
     def resample(self, indexer=None, skipna=None, closed=None, label=None,
-                 base=0, keep_attrs=None, **indexer_kwargs):
+                 base=0, keep_attrs=None, loffset=None, **indexer_kwargs):
         """Returns a Resample object for performing resampling operations.
 
         Handles both downsampling and upsampling. If any intervals contain no
@@ -612,6 +612,9 @@ def resample(self, indexer=None, skipna=None, closed=None, label=None,
             For frequencies that evenly subdivide 1 day, the "origin" of the
             aggregated intervals. For example, for '24H' frequency, base could
             range from 0 through 23.
+        loffset : timedelta or str, optional
+            Offset used to adjust the resampled time labels. Some pandas date
+            offset strings are supported.
         keep_attrs : bool, optional
             If True, the object's attributes (`attrs`) will be copied from
             the original object to the new one.  If False (default), the new
@@ -700,7 +703,9 @@ def resample(self, indexer=None, skipna=None, closed=None, label=None,
 
         group = DataArray(dim_coord, coords=dim_coord.coords,
                           dims=dim_coord.dims, name=RESAMPLE_DIM)
-        grouper = pd.Grouper(freq=freq, closed=closed, label=label, base=base)
+        # TODO: to_offset() call required for pandas==0.19.2
+        grouper = pd.Grouper(freq=freq, closed=closed, label=label, base=base,
+                             loffset=pd.tseries.frequencies.to_offset(loffset))
         resampler = self._resample_cls(self, group=group, dim=dim_name,
                                        grouper=grouper,
                                        resample_dim=RESAMPLE_DIM)

xarray/core/groupby.py

@@ -3,6 +3,7 @@
 import functools
 import warnings
 
+import datetime
 import numpy as np
 import pandas as pd
 
@@ -154,6 +155,32 @@ def _unique_and_monotonic(group):
         return index.is_unique and index.is_monotonic
 
 
+def _apply_loffset(grouper, result):
+    """
+    (copied from pandas)
+    if loffset is set, offset the result index
+
+    This is NOT an idempotent routine, it will be applied
+    exactly once to the result.
+
+    Parameters
+    ----------
+    result : Series or DataFrame
+        the result of resample
+    """
+
+    needs_offset = (
+        isinstance(grouper.loffset, (pd.DateOffset, datetime.timedelta))
+        and isinstance(result.index, pd.DatetimeIndex)
+        and len(result.index) > 0
+    )
+
+    if needs_offset:
+        result.index = result.index + grouper.loffset
+
+    grouper.loffset = None
+
+
 class GroupBy(SupportsArithmetic):
     """A object that implements the split-apply-combine pattern.
 
@@ -235,6 +262,7 @@ def __init__(self, obj, group, squeeze=False, grouper=None, bins=None,
                 raise ValueError('index must be monotonic for resampling')
             s = pd.Series(np.arange(index.size), index)
             first_items = s.groupby(grouper).first()
+            _apply_loffset(grouper, first_items)
             full_index = first_items.index
             if first_items.isnull().any():
                 first_items = first_items.dropna()

xarray/tests/test_dataarray.py

@@ -2273,6 +2273,11 @@ def test_resample(self):
         actual = array.resample(time='24H').reduce(np.mean)
         assert_identical(expected, actual)
 
+        actual = array.resample(time='24H', loffset='-12H').mean()
+        expected = DataArray(array.to_series().resample('24H', loffset='-12H')
+                             .mean())
+        assert_identical(expected, actual)
+
         with raises_regex(ValueError, 'index must be monotonic'):
             array[[2, 0, 1]].resample(time='1D')
 

xarray/tests/test_dataset.py

@@ -2804,6 +2804,18 @@ def test_resample_by_mean_with_keep_attrs(self):
         expected = ds.attrs
         assert expected == actual
 
+    def test_resample_loffset(self):
+        times = pd.date_range('2000-01-01', freq='6H', periods=10)
+        ds = Dataset({'foo': (['time', 'x', 'y'], np.random.randn(10, 5, 3)),
+                      'bar': ('time', np.random.randn(10), {'meta': 'data'}),
+                      'time': times})
+        ds.attrs['dsmeta'] = 'dsdata'
+
+        actual = ds.resample(time='24H', loffset='-12H').mean('time').time
+        expected = xr.DataArray(ds.bar.to_series()
+                                .resample('24H', loffset='-12H').mean()).time
+        assert_identical(expected, actual)
+
     def test_resample_by_mean_discarding_attrs(self):
         times = pd.date_range('2000-01-01', freq='6H', periods=10)
         ds = Dataset({'foo': (['time', 'x', 'y'], np.random.randn(10, 5, 3)),