Fixes to DatetimeConverter speed up following PR comments and corresponding note in release notes..

Author: agijsberts

doc/source/release.rst

@@ -159,6 +159,8 @@ Improvements to existing features
 - ``StataWriter`` and ``DataFrame.to_stata`` accept time stamp and data labels (:issue:`6545`)
 - offset/freq info now in Timestamp __repr__ (:issue:`4553`)
 - Support passing ``encoding`` with xlwt (:issue:`3710`)
+- Performance improvement when converting ``DatetimeIndex`` to floating ordinals 
+  using ``DatetimeConverter`` (:issue:`6636`)
 
 .. _release.bug_fixes-0.14.0:
 

pandas/tseries/converter.py

@@ -16,6 +16,7 @@
 import pandas.core.common as com
 from pandas.core.index import Index
 
+from pandas import Series
 from pandas.tseries.index import date_range
 import pandas.tseries.tools as tools
 import pandas.tseries.frequencies as frequencies
@@ -144,8 +145,8 @@ def _dt_to_float_ordinal(dt):
     preserving hours, minutes, seconds and microseconds.  Return value
     is a :func:`float`.
     """
-    if isinstance(dt, np.ndarray) and com.is_datetime64_ns_dtype(dt):
-        base = dates.epoch2num(dt.astype(np.int64) / 1.0E9)
+    if isinstance(dt, (np.ndarray, Series)) and com.is_datetime64_ns_dtype(dt):
+        base = dates.epoch2num(dt.asi8 / 1.0E9)
     else:
         base = dates.date2num(dt)
     return base