* _iotools.StringConverter

  - prevents a `default` of 0 to be overwritten during initialization
  - allows the `missing_values` to be a comma-separated string
* io.genfromtxt
  - `usecols` can now be a single integer
  - for `usecols` and `names` to list (for compatibility w/ Python 2.5)
  - negative values in `usecols` are properly transformed to positive integers
  - fixed `usecols` with named columns

1 files changed, 10 insertions, 1 deletions

diff --git a/numpy/lib/tests/test__iotools.py b/numpy/lib/tests/test__iotools.py
index ed0f4dc63..11ce4047b 100644
--- a/numpy/lib/tests/test__iotools.py
+++ b/numpy/lib/tests/test__iotools.py
@@ -191,7 +191,16 @@ class TestStringConverter(TestCase):
         converter.upgrade('3.14159265')
         assert_equal(converter.default, 0)
         assert_equal(converter.type, np.dtype(float))
-
+    #
+    def test_keep_default_zero(self):
+        "Check that we don't lose a default of 0"
+        converter = StringConverter(int, default=0, missing_values="N/A")
+        assert_equal(converter.default, 0)
+    #
+    def test_keep_missing_values(self):
+        "Check that we're not losing missing values"
+        converter = StringConverter(int, default=0, missing_values="N/A")
+        assert_equal(converter.missing_values, set(['', 'N/A']))
 
 #-------------------------------------------------------------------------------