simplify timestamp precision issue

Author: reidy-p

doc/source/whatsnew/v0.23.0.txt

@@ -780,12 +780,12 @@ Note that this change also applies to :meth:`DataFrame.append`, which has also r
 .. _whatsnew_0230.api_breaking.end_time:
 
 Time values in ``dt.end_time`` and ``to_timestamp(how='end')``
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
-The time values in ``Period`` and ``PeriodIndex`` objects are now adjusted
-appropriately when calling :attr:`Series.dt.end_time`, :attr:`Period.end_time`,
+The time values in :class:`Period` and :class:`PeriodIndex` objects are now set
+to '23:59:59.999999999' when calling :attr:`Series.dt.end_time`, :attr:`Period.end_time`,
 :attr:`PeriodIndex.end_time`, :func:`Period.to_timestamp()` with ``how='end'``,
-and :func:`PeriodIndex.to_timestamp()` with ``how='end'`` (:issue:`17157`)
+or :func:`PeriodIndex.to_timestamp()` with ``how='end'`` (:issue:`17157`)
 
 Previous Behavior:
 
@@ -797,22 +797,11 @@ Previous Behavior:
    In [4]: pd.Series(p).dt.end_time[0]
    Out[4]: Timestamp(2017-01-01 00:00:00)
 
-   In [5]: pd.Series(pi).dt.end_time[0]
+   In [5]: pi.end_time[0]
    Out[5]: Timestamp(2017-01-01 00:00:00)
 
-   In [6]: p.end_time
-   Out[6]: Timestamp(2017-01-01 23:59:59.999999999)
-
-   In [7]: pi.end_time[0]
-   Out[7]: Timestamp(2017-01-01 00:00:00)
-
-   In [8]: p.to_timestamp(how='end')
-   Out[8]: Timestamp(2017-01-01 00:00:00)
-
-   In [9]: pi.to_timestamp(how='end')[0]
-   Out[9]: Timestamp(2017-01-01 00:00:00)
-
-Current Behavior
+   In [6]: pi.to_timestamp(how='end')
+   Out[6]: Timestamp(2017-01-01 00:00:00)
 
 .. ipython:: python
 
@@ -821,15 +810,9 @@ Current Behavior
 
    pd.Series(p).dt.end_time[0]
 
-   pd.Series(pi).dt.end_time[0]
-
-   p.end_time
-
    pi.end_time[0]
 
-   p.to_timestamp(how='end')
-
-   pi.to_timestamp(how='end')[0]
+   pi.to_timestamp(how='end')
 
 
 Build Changes

pandas/core/indexes/datetimes.py

@@ -1566,6 +1566,47 @@ def to_julian_date(self):
 DatetimeIndex._add_datetimelike_methods()
 
 
+<<<<<<< 52e0171b2086c87362ae33db0d3fe9cc2c4afc2a
+=======
+def _generate_regular_range(cls, start, end, periods, freq):
+    if isinstance(freq, Tick):
+        stride = freq.nanos
+        if periods is None:
+            b = Timestamp(start).value
+            # cannot just use e = Timestamp(end) + 1 because arange breaks when
+            # stride is too large, see GH10887
+            e = (b + (Timestamp(end).value - b) // stride * stride +
+                 stride // 2 + 1)
+            # end.tz == start.tz by this point due to _generate implementation
+            tz = start.tz
+        elif start is not None:
+            b = Timestamp(start).value
+            e = b + np.int64(periods) * stride
+            tz = start.tz
+        elif end is not None:
+            e = Timestamp(end).value + stride
+            b = e - np.int64(periods) * stride
+            tz = end.tz
+        else:
+            raise ValueError("at least 'start' or 'end' should be specified "
+                             "if a 'period' is given.")
+
+        data = np.arange(b, e, stride, dtype=np.int64)
+        # TODO: Do we need to use _simple_new here?  just return data.view?
+        data = cls._simple_new(data.view(_NS_DTYPE), None, tz=tz)
+    else:
+
+        xdr = generate_range(start=start, end=end,
+                             periods=periods, offset=offset)
+
+        dates = list(xdr)
+        # utc = len(dates) > 0 and dates[0].tzinfo is not None
+        data = tools.to_datetime(dates)
+
+    return data
+
+
+>>>>>>> simplify timestamp precision issue
 def date_range(start=None, end=None, periods=None, freq=None, tz=None,
                normalize=False, name=None, closed=None, **kwargs):
     """

pandas/core/indexes/period.py

@@ -519,19 +519,7 @@ def to_timestamp(self, freq=None, how='start'):
         base, mult = _gfc(freq)
         new_data = self.asfreq(freq, how)
 
-        end = how == 'E'
-        if end:
-            indexer = np.where(new_data.notnull())
-            # move forward one period
-            new_data._values[indexer] += 1
-            ndarray_vals = new_data._ndarray_values
-            new_data = period.periodarr_to_dt64arr(ndarray_vals, base)
-            # subtract one nanosecond
-            new_data[indexer] -= 1
-        else:
-            ndarray_vals = new_data._ndarray_values
-            new_data = period.periodarr_to_dt64arr(ndarray_vals, base)
-
+        new_data = period.periodarr_to_dt64arr(new_data._ndarray_values, base)
         return DatetimeIndex(new_data, freq='infer', name=self.name)
 
     @property

pandas/tests/indexes/period/test_period.py

@@ -367,11 +367,13 @@ def test_periods_number_check(self):
             period_range('2011-1-1', '2012-1-1', 'B')
 
     def test_start_time(self):
+        # GH 17157
         index = PeriodIndex(freq='M', start='2016-01-01', end='2016-05-31')
         expected_index = date_range('2016-01-01', end='2016-05-31', freq='MS')
         tm.assert_index_equal(index.start_time, expected_index)
 
     def test_end_time(self):
+        # GH 17157
         index = PeriodIndex(freq='M', start='2016-01-01', end='2016-05-31')
         expected_index = date_range('2016-01-01', end='2016-05-31', freq='M')
         expected_index = expected_index.shift(1, freq='D').shift(-1, freq='ns')