Merge commit 'v0.9.0rc1-7-g8c0e206' into debian

* commit 'v0.9.0rc1-7-g8c0e206':
  BUG: fix Cython tz_convert bug with time zones that haven't had a UTC transition in a long time. close pandas-dev#1946
  BUG: fix buglet
  BUG: try fixing tzlocal bug
  Minor fixes to time series doc.
  Adding DataFrame methods to API reference.
  Added Series functions to API doc.
  BUG: fix segfault in SeriesGrouper with non-contiguous index

Author: yarikoptic

RELEASE.rst

@@ -203,6 +203,8 @@ pandas 0.9.0
     partial-indexing a hierarchically-indexed DataFrame (#1796)
   - Support multiple column selection in DataFrame.__getitem__ with duplicate
     columns (#1943)
+  - Fix time zone localization bug causing improper fields (e.g. hours) in time
+    zones that have not had a UTC transition in a long time (#1946)
 
 pandas 0.8.1
 ============

doc/source/api.rst

@@ -165,6 +165,7 @@ Binary operator functions
    Series.sub
    Series.combine
    Series.combine_first
+   Series.round
 
 Function application, GroupBy
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -182,25 +183,37 @@ Computations / Descriptive Stats
 .. autosummary::
    :toctree: generated/
 
+   Series.abs
+   Series.any
    Series.autocorr
+   Series.between
    Series.clip
    Series.clip_lower
    Series.clip_upper
    Series.corr
    Series.count
+   Series.cov
+   Series.cummax
+   Series.cummin
    Series.cumprod
    Series.cumsum
    Series.describe
    Series.diff
+   Series.kurt
+   Series.mad
    Series.max
    Series.mean
    Series.median
    Series.min
+   Series.nunique
+   Series.pct_change
    Series.prod
    Series.quantile
+   Series.rank
    Series.skew
    Series.std
    Series.sum
+   Series.unique
    Series.var
    Series.value_counts
 
@@ -211,11 +224,19 @@ Reindexing / Selection / Label manipulation
 
    Series.align
    Series.drop
+   Series.first
+   Series.head
+   Series.idxmax
+   Series.idxmin
+   Series.isin
+   Series.last
    Series.reindex
    Series.reindex_like
    Series.rename
+   Series.reset_index
    Series.select
    Series.take
+   Series.tail
    Series.truncate
 
 Missing data handling
@@ -234,9 +255,11 @@ Reshaping, sorting
 
    Series.argsort
    Series.order
+   Series.reorder_levels
    Series.sort
    Series.sort_index
    Series.sortlevel
+   Series.swaplevel
    Series.unstack
 
 Combining / joining / merging
@@ -245,6 +268,8 @@ Combining / joining / merging
    :toctree: generated/
 
    Series.append
+   Series.replace
+   Series.update
 
 Time series-related
 ~~~~~~~~~~~~~~~~~~~
@@ -257,6 +282,9 @@ Time series-related
    Series.first_valid_index
    Series.last_valid_index
    Series.weekday
+   Series.resample
+   Series.tz_convert
+   Series.tz_localize
 
 Plotting
 ~~~~~~~~
@@ -277,6 +305,7 @@ Serialization / IO / Conversion
    Series.to_csv
    Series.to_dict
    Series.to_sparse
+   Series.to_string
 
 .. _api.dataframe:
 
@@ -308,18 +337,24 @@ Conversion / Constructors
 
    DataFrame.__init__
    DataFrame.astype
+   DataFrame.convert_objects
    DataFrame.copy
 
 Indexing, iteration
 ~~~~~~~~~~~~~~~~~~~
 .. autosummary::
    :toctree: generated/
 
+   DataFrame.head
    DataFrame.ix
    DataFrame.insert
    DataFrame.__iter__
    DataFrame.iteritems
+   DataFrame.iterrows
+   DataFrame.itertuples
+   DataFrame.lookup
    DataFrame.pop
+   DataFrame.tail
    DataFrame.xs
 
 Binary operator functions
@@ -356,23 +391,31 @@ Computations / Descriptive Stats
 .. autosummary::
    :toctree: generated/
 
+   DataFrame.abs
+   DataFrame.any
    DataFrame.clip
    DataFrame.clip_lower
    DataFrame.clip_upper
    DataFrame.corr
    DataFrame.corrwith
    DataFrame.count
+   DataFrame.cov
+   DataFrame.cummax
+   DataFrame.cummin
    DataFrame.cumprod
    DataFrame.cumsum
    DataFrame.describe
    DataFrame.diff
+   DataFrame.kurt
    DataFrame.mad
    DataFrame.max
    DataFrame.mean
    DataFrame.median
    DataFrame.min
+   DataFrame.pct_change
    DataFrame.prod
    DataFrame.quantile
+   DataFrame.rank
    DataFrame.skew
    DataFrame.sum
    DataFrame.std
@@ -387,15 +430,24 @@ Reindexing / Selection / Label manipulation
    DataFrame.add_suffix
    DataFrame.align
    DataFrame.drop
+   DataFrame.drop_duplicates
+   DataFrame.duplicated
    DataFrame.filter
+   DataFrame.first
+   DataFrame.head
+   DataFrame.idxmax
+   DataFrame.idxmin
+   DataFrame.last
    DataFrame.reindex
+   DataFrame.reindex_axis
    DataFrame.reindex_like
    DataFrame.rename
+   DataFrame.reset_index
    DataFrame.select
+   DataFrame.set_index
+   DataFrame.tail
    DataFrame.take
    DataFrame.truncate
-   DataFrame.head
-   DataFrame.tail
 
 .. _api.dataframe.missing:
 
@@ -412,24 +464,29 @@ Reshaping, sorting, transposing
 .. autosummary::
    :toctree: generated/
 
-   DataFrame.sort_index
    DataFrame.delevel
    DataFrame.pivot
+   DataFrame.reorder_levels
+   DataFrame.sort
+   DataFrame.sort_index
    DataFrame.sortlevel
    DataFrame.swaplevel
    DataFrame.stack
    DataFrame.unstack
    DataFrame.T
+   DataFrame.to_panel
    DataFrame.transpose
 
 Combining / joining / merging
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 .. autosummary::
    :toctree: generated/
 
+   DataFrame.append
    DataFrame.join
    DataFrame.merge
-   DataFrame.append
+   DataFrame.replace
+   DataFrame.update
 
 Time series-related
 ~~~~~~~~~~~~~~~~~~~
@@ -440,12 +497,18 @@ Time series-related
    DataFrame.shift
    DataFrame.first_valid_index
    DataFrame.last_valid_index
+   DataFrame.resample
+   DataFrame.to_period
+   DataFrame.to_timestamp
+   DataFrame.tz_convert
+   DataFrame.tz_localize
 
 Plotting
 ~~~~~~~~
 .. autosummary::
    :toctree: generated/
 
+   DataFrame.boxplot
    DataFrame.hist
    DataFrame.plot
 
@@ -455,16 +518,19 @@ Serialization / IO / Conversion
    :toctree: generated/
 
    DataFrame.from_csv
+   DataFrame.from_dict
+   DataFrame.from_items
    DataFrame.from_records
+   DataFrame.info
+   DataFrame.load
+   DataFrame.save
    DataFrame.to_csv
-   DataFrame.to_excel
    DataFrame.to_dict
+   DataFrame.to_excel
+   DataFrame.to_html
    DataFrame.to_records
    DataFrame.to_sparse
    DataFrame.to_string
-   DataFrame.save
-   DataFrame.load
-   DataFrame.info
 
 .. _api.panel:
 

doc/source/timeseries.rst

@@ -140,7 +140,7 @@ to create timestamp indexes.
    index = bdate_range('2012-1-1', periods=250)
    index
 
-Convenience functions like ``date_range`` and ``bdate_range`` utilizes a
+Convenience functions like ``date_range`` and ``bdate_range`` utilize a
 variety of frequency aliases. The default frequency for ``date_range`` is a
 **calendar day** while the default for ``bdate_range`` is a **business day**
 
@@ -156,7 +156,7 @@ variety of frequency aliases. The default frequency for ``date_range`` is a
    rng
 
 ``date_range`` and ``bdate_range`` makes it easy to generate a range of dates
-using various combinations of its parameters like ``start``, ``end``,
+using various combinations of parameters like ``start``, ``end``,
 ``periods``, and ``freq``:
 
 .. ipython:: python
@@ -586,7 +586,7 @@ different parameters to control the frequency conversion and resampling
 operation.
 
 The ``how`` parameter can be a function name or numpy array function that takes
-and array and produces an aggregated values:
+an array and produces aggregated values:
 
 .. ipython:: python
 

pandas/src/datetime.pyx

@@ -527,11 +527,11 @@ cdef class _TSObject:
         def __get__(self):
             return self.value
 
-cpdef _get_utcoffset(tzinfo):
+cpdef _get_utcoffset(tzinfo, obj):
     try:
         return tzinfo._utcoffset
     except AttributeError:
-        return tzinfo.utcoffset(None)
+        return tzinfo.utcoffset(obj)
 
 # helper to extract datetime and int64 from several different possibilities
 cpdef convert_to_tsobject(object ts, object tz=None):
@@ -570,20 +570,22 @@ cpdef convert_to_tsobject(object ts, object tz=None):
                 ts = tz.normalize(ts)
                 obj.value = _pydatetime_to_dts(ts, &obj.dts)
                 obj.tzinfo = ts.tzinfo
-            elif tz is not pytz.utc:
+            elif not _is_utc(tz):
                 ts = tz.localize(ts)
                 obj.value = _pydatetime_to_dts(ts, &obj.dts)
-                obj.value -= _delta_to_nanoseconds(_get_utcoffset(ts.tzinfo))
+                offset = _get_utcoffset(ts.tzinfo, ts)
+                obj.value -= _delta_to_nanoseconds(offset)
                 obj.tzinfo = ts.tzinfo
             else:
                 # UTC
                 obj.value = _pydatetime_to_dts(ts, &obj.dts)
-                obj.tzinfo = tz
+                obj.tzinfo = pytz.utc
         else:
             obj.value = _pydatetime_to_dts(ts, &obj.dts)
             obj.tzinfo = ts.tzinfo
             if obj.tzinfo is not None and not _is_utc(obj.tzinfo):
-                obj.value -= _delta_to_nanoseconds(_get_utcoffset(obj.tzinfo))
+                offset = _get_utcoffset(obj.tzinfo, ts)
+                obj.value -= _delta_to_nanoseconds(offset)
         _check_dts_bounds(obj.value, &obj.dts)
         return obj
     elif PyDate_Check(ts):
@@ -946,7 +948,7 @@ def tz_convert(ndarray[int64_t] vals, object tz1, object tz2):
     if pos == 0:
         raise ValueError('First time before start of DST info')
     elif pos == len(trans):
-        return result + deltas[-1]
+        return utc_dates + deltas[-1]
 
     # TODO: this assumed sortedness :/
     pos -= 1
@@ -1015,6 +1017,11 @@ def _get_transitions(tz):
         if hasattr(tz, '_utc_transition_times'):
             arr = np.array(tz._utc_transition_times, dtype='M8[ns]')
             arr = arr.view('i8')
+            try:
+                if tz._utc_transition_times[0].year == 1:
+                    arr[0] = NPY_NAT + 1
+            except Exception:
+                pass
         else:
             arr = np.array([NPY_NAT + 1], dtype=np.int64)
         trans_cache[tz] = arr
@@ -1028,15 +1035,15 @@ def _get_deltas(tz):
         # tzoffset not hashable in Python 3
         hash(tz)
     except TypeError:
-        num = int(total_seconds(_get_utcoffset(tz))) * 1000000000
+        num = int(total_seconds(_get_utcoffset(tz, None))) * 1000000000
         return np.array([num], dtype=np.int64)
 
     if tz not in utc_offset_cache:
         if hasattr(tz, '_utc_transition_times'):
             utc_offset_cache[tz] = _unbox_utcoffsets(tz._transition_info)
         else:
             # static tzinfo
-            num = int(total_seconds(_get_utcoffset(tz))) * 1000000000
+            num = int(total_seconds(_get_utcoffset(tz, None))) * 1000000000
             utc_offset_cache[tz] = np.array([num], dtype=np.int64)
     return utc_offset_cache[tz]