BUG: indexing error when setting item in empty Series which has a frequency

[rekcahpassyla]

Another somewhat degenerate case that works in 0.15.2, fails in 0.16.1

import pandas as pd

pd.show_versions()


ts = pd.TimeSeries()

ts[pd.datetime(2012, 1, 1)] = 47


ts2 = pd.TimeSeries(0, pd.date_range('2011-01-01', '2011-01-01'))[:0]

ts2[pd.datetime(2012, 1, 1)] = 47

0.15.2 output:

INSTALLED VERSIONS
------------------
commit: None
python: 2.7.9.final.0
python-bits: 64
OS: Windows
OS-release: 7
machine: AMD64
processor: Intel64 Family 6 Model 26 Stepping 5, GenuineIntel
byteorder: little
LC_ALL: None
LANG: None

pandas: 0.15.2
nose: 1.3.4
Cython: 0.21
numpy: 1.9.2
scipy: 0.14.0
statsmodels: 0.5.0
IPython: 3.0.0-dev
sphinx: 1.2.3
patsy: 0.3.0
dateutil: 2.4.2
pytz: 2015.4
bottleneck: 0.8.0
tables: 3.1.1
numexpr: 2.3.1
matplotlib: 1.4.2
openpyxl: None
xlrd: 0.9.3
xlwt: None
xlsxwriter: 0.5.7
lxml: 3.4.0
bs4: 4.3.2
html5lib: 0.999
httplib2: None
apiclient: None
rpy2: None
sqlalchemy: 0.9.7
pymysql: None
psycopg2: None

0.16.1 output:

INSTALLED VERSIONS
------------------
commit: None
python: 2.7.9.final.0
python-bits: 64
OS: Windows
OS-release: 7
machine: AMD64
processor: Intel64 Family 6 Model 26 Stepping 5, GenuineIntel
byteorder: little
LC_ALL: None
LANG: None

pandas: 0.16.1-46-g0aceb38
nose: 1.3.6
Cython: 0.22
numpy: 1.9.2
scipy: 0.14.0
statsmodels: 0.6.1
IPython: 3.1.0
sphinx: 1.3.1
patsy: 0.3.0
dateutil: 2.4.2
pytz: 2015.4
bottleneck: 0.8.0
tables: 3.1.1
numexpr: 2.3.1
matplotlib: 1.4.3
openpyxl: None
xlrd: 0.9.3
xlwt: None
xlsxwriter: 0.7.2
lxml: None
bs4: 4.3.2
html5lib: 0.999
httplib2: None
apiclient: None
sqlalchemy: 1.0.4
pymysql: None
psycopg2: None

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
c:\test_set_empty_series_with_freq.py in <module>()
     11 ts2 = pd.TimeSeries(0, pd.date_range('2011-01-01', '2011-01-01'))[:0]
     12
---> 13 ts2[pd.datetime(2012, 1, 1)] = 47

c:\python\envs\pandas-0.16.1\lib\site-packages\pandas\core\series.pyc in __setitem__(self, key, value)
    687         # do the setitem
    688         cacher_needs_updating = self._check_is_chained_assignment_possible()
--> 689         setitem(key, value)
    690         if cacher_needs_updating:
    691             self._maybe_update_cacher()

c:\python\envs\pandas-0.16.1\lib\site-packages\pandas\core\series.pyc in setitem(key, value)
    660                             pass
    661                 try:
--> 662                     self.loc[key] = value
    663                 except:
    664                     print ""

c:\python\envs\pandas-0.16.1\lib\site-packages\pandas\core\indexing.pyc in __setitem__(self, key, value)
    113     def __setitem__(self, key, value):
    114         indexer = self._get_setitem_indexer(key)
--> 115         self._setitem_with_indexer(indexer, value)
    116
    117     def _has_valid_type(self, k, axis):

c:\python\envs\pandas-0.16.1\lib\site-packages\pandas\core\indexing.pyc in _setitem_with_indexer(self, indexer, value)
    272                 if self.ndim == 1:
    273                     index = self.obj.index
--> 274                     new_index = index.insert(len(index),indexer)
    275
    276                     # this preserves dtype of the value

c:\python\envs\pandas-0.16.1\lib\site-packages\pandas\tseries\index.pyc in insert(self, loc, item)
   1523             # check freq can be preserved on edge cases
   1524             if self.freq is not None:
-> 1525                 if (loc == 0 or loc == -len(self)) and item + self.freq == self[0]:
   1526                     freq = self.freq
   1527                 elif (loc == len(self)) and item - self.freq == self[-1]:

c:\python\envs\pandas-0.16.1\lib\site-packages\pandas\tseries\index.pyc in __getitem__(self, key)
   1351         getitem = self._data.__getitem__
   1352         if np.isscalar(key):
-> 1353             val = getitem(key)
   1354             return Timestamp(val, offset=self.offset, tz=self.tz)
   1355         else:

IndexError: index 0 is out of bounds for axis 0 with size 0

Proposed #10194

[jreback]

This actually is a bit tricker. You are slicing another object, then trying to set it. This actually should hit the SettingWIthCopyWarning path, though it should actually work.

Side note is that we don't use TimeSeries anymore (its just a synonym for Series and will be deprecated in 0.17.0)

In [2]: ts2 = Series(0, pd.date_range('2011-01-01', '2011-01-01')) 

In [3]: ts2
Out[3]: 
2011-01-01    0
Freq: D, dtype: int64

In [4]: ts3 = ts2[:0]

In [5]: ts3
Out[5]: Series([], Freq: D, dtype: int64)

# this is the original structure
In [8]: ts3._data.blocks[0].values.base
Out[8]: array([0])

[rekcahpassyla]

Actually, I've just found that setting an item on a newly created empty series with frequency still returns an error. Since it's a new series, there is no view involved.

In [142]: pd.show_versions()

INSTALLED VERSIONS
------------------
commit: None
python: 2.7.10.final.0
python-bits: 64
OS: Windows
OS-release: 7
machine: AMD64
processor: Intel64 Family 6 Model 26 Stepping 5, GenuineIntel
byteorder: little
LC_ALL: None
LANG: None

pandas: 0.16.2
nose: 1.3.4
Cython: 0.22
numpy: 1.9.2
scipy: 0.14.0
statsmodels: 0.6.1
IPython: 3.0.0
sphinx: 1.3
patsy: 0.3.0
dateutil: 2.4.2
pytz: 2015.4
bottleneck: 0.8.0
tables: 3.1.1
numexpr: 2.3.1
matplotlib: 1.4.3
openpyxl: 1.8.5
xlrd: 0.9.3
xlwt: 0.7.5
xlsxwriter: 0.6.7
lxml: 3.4.2
bs4: 4.3.2
html5lib: 0.999
httplib2: None
apiclient: None
sqlalchemy: 0.9.9
pymysql: None
psycopg2: None

In [143]: series = pd.Series([], freq='D')
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-143-123100163557> in <module>()
----> 1 series = pd.Series([], freq='D')

TypeError: __init__() got an unexpected keyword argument 'freq'

In [144]: series = pd.Series([], pd.DatetimeIndex([], freq='D'))

In [145]: dt = pd.Timestamp('2011-01-01')

In [146]: series.loc[dt] = 9
---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-146-c577a9a95757> in <module>()
----> 1 series.loc[dt] = 9

C:\dev\bin\Anaconda\envs\pandas-0.16.2\lib\site-packages\pandas-0.16.2-py2.7-win-amd64.egg\pandas\core\indexing.pyc in __setitem__(self, key, value)
    113     def __setitem__(self, key, value):
    114         indexer = self._get_setitem_indexer(key)
--> 115         self._setitem_with_indexer(indexer, value)
    116
    117     def _has_valid_type(self, k, axis):

C:\dev\bin\Anaconda\envs\pandas-0.16.2\lib\site-packages\pandas-0.16.2-py2.7-win-amd64.egg\pandas\core\indexing.pyc in _setitem_with_indexer(self, indexer, value)
    281                 if self.ndim == 1:
    282                     index = self.obj.index
--> 283                     new_index = index.insert(len(index),indexer)
    284
    285                     # this preserves dtype of the value

C:\dev\bin\Anaconda\envs\pandas-0.16.2\lib\site-packages\pandas-0.16.2-py2.7-win-amd64.egg\pandas\tseries\index.pyc in insert(self, loc, item)
   1499             # check freq can be preserved on edge cases
   1500             if self.freq is not None:
-> 1501                 if (loc == 0 or loc == -len(self)) and item + self.freq == self[0]:
   1502                     freq = self.freq
   1503                 elif (loc == len(self)) and item - self.freq == self[-1]:

C:\dev\bin\Anaconda\envs\pandas-0.16.2\lib\site-packages\pandas-0.16.2-py2.7-win-amd64.egg\pandas\tseries\base.pyc in __getitem__(self, key)
     73         getitem = self._data.__getitem__
     74         if np.isscalar(key):
---> 75             val = getitem(key)
     76             return self._box_func(val)
     77         else:

IndexError: index 0 is out of bounds for axis 0 with size 0


In [148]: series_nofreq = pd.Series([], pd.DatetimeIndex([]))

In [149]: series_nofreq.loc[pd.Timestamp('2011-01-01')] = 47

In [150]: