summaryrefslogtreecommitdiff
path: root/numpy/core/defchararray.py
diff options
context:
space:
mode:
Diffstat (limited to 'numpy/core/defchararray.py')
-rw-r--r--numpy/core/defchararray.py14
1 files changed, 7 insertions, 7 deletions
diff --git a/numpy/core/defchararray.py b/numpy/core/defchararray.py
index 201a860cc..9e9e78122 100644
--- a/numpy/core/defchararray.py
+++ b/numpy/core/defchararray.py
@@ -1,4 +1,4 @@
-from numerictypes import string, unicode_, integer, object_
+from numerictypes import string_, unicode_, integer, object_
from numeric import ndarray, broadcast, empty
from numeric import array as narray
import sys
@@ -8,7 +8,7 @@ __all__ = ['chararray']
_globalvar = 0
_unicode = unicode
-# special sub-class for character arrays (string and unicode_)
+# special sub-class for character arrays (string_ and unicode_)
# This adds equality testing and methods of str and unicode types
# which operate on an element-by-element basis
@@ -21,7 +21,7 @@ class chararray(ndarray):
if unicode:
dtype = unicode_
else:
- dtype = string
+ dtype = string_
_globalvar = 1
if buffer is None:
@@ -300,17 +300,17 @@ def array(obj, itemsize=None, copy=True, unicode=False, order=None):
itemsize = obj.itemsize
if copy or (itemsize != obj.itemsize) \
or (not unicode and obj.dtype == unicode_) \
- or (unicode and obj.dtype == string):
+ or (unicode and obj.dtype == string_):
return obj.astype("%s%d" % (obj.dtype.char, itemsize))
else:
return obj
- if isinstance(obj, ndarray) and (obj.dtype in [unicode_, string]):
+ if isinstance(obj, ndarray) and (obj.dtype in [unicode_, string_]):
new = obj.view(chararray)
- if unicode and obj.dtype == string:
+ if unicode and obj.dtype == string_:
return new.astype((unicode_, obj.itemsize))
elif obj.dtype == unicode_:
- return new.astype((string, obj.itemsize))
+ return new.astype((string_, obj.itemsize))
if copy: return new.copy()
else: return new