Add -NaN and -nan to default NA values

Author: Caleb Epstein

doc/source/io.rst

@@ -575,7 +575,7 @@ the corresponding equivalent values will also imply a missing value (in this cas
 
 To completely override the default values that are recognized as missing, specify ``keep_default_na=False``.
 The default ``NaN`` recognized values are ``['-1.#IND', '1.#QNAN', '1.#IND', '-1.#QNAN', '#N/A','N/A', 'NA',
-'#NA', 'NULL', 'NaN', 'nan']``.
+'#NA', 'NULL', 'NaN', '-NaN', 'nan', '-nan']``.
 
 .. code-block:: python
 

doc/source/release.rst

@@ -64,6 +64,8 @@ API Changes
   - ``autocorrelation_plot`` now accepts ``**kwargs``. (:issue:`5623`)
   - ``convert_objects`` now accepts a ``convert_timedeltas='coerce'`` argument to allow forced dtype conversion of
     timedeltas (:issue:`5458`,:issue:`5689`)
+  - Add ``-NaN`` and ``-nan`` to the default set of NA values
+    (:issue:`5952`).  See :ref:`NA Values <io.na_values>`.
 
 Experimental Features
 ~~~~~~~~~~~~~~~~~~~~~

doc/source/v0.13.1.txt

@@ -18,6 +18,9 @@ There are several new or updated docs sections including:
 API changes
 ~~~~~~~~~~~
 
+- Add ``-NaN`` and ``-nan`` to the default set of NA values (:issue:`5952`).
+  See :ref:`NA Values <io.na_values>`.
+
 Prior Version Deprecations/Changes
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

pandas/io/parsers.py

@@ -449,7 +449,7 @@ def read_fwf(filepath_or_buffer, colspecs='infer', widths=None, **kwds):
 # '1.#INF','-1.#INF', '1.#INF000000',
 _NA_VALUES = set([
     '-1.#IND', '1.#QNAN', '1.#IND', '-1.#QNAN', '#N/A', 'N/A', 'NA', '#NA',
-    'NULL', 'NaN', 'nan', ''
+    'NULL', 'NaN', '-NaN', 'nan', '-nan', ''
 ])
 
 

pandas/io/tests/test_parsers.py

@@ -33,6 +33,7 @@
 import pandas.tseries.tools as tools
 
 from numpy.testing.decorators import slow
+from numpy.testing import assert_array_equal
 
 from pandas.parser import OverflowError
 
@@ -685,8 +686,8 @@ def test_non_string_na_values(self):
     def test_default_na_values(self):
         _NA_VALUES = set(['-1.#IND', '1.#QNAN', '1.#IND', '-1.#QNAN',
                           '#N/A','N/A', 'NA', '#NA', 'NULL', 'NaN',
-                          'nan', ''])
-
+                          'nan', '-NaN', '-nan', ''])
+        assert_array_equal (_NA_VALUES, parsers._NA_VALUES)
         nv = len(_NA_VALUES)
         def f(i, v):
             if i == 0: