[Numpy-discussion] How to distinguish between number and string dypes

[Bruce Southey]

On 05/27/2010 10:40 AM, Vincent Davis wrote:
> On Thu, May 27, 2010 at 8:39 AM, Keith Goodman <kwgoodman at gmail.com 
> <mailto:kwgoodman at gmail.com>> wrote:
>
>     To see if it is a number could you use something like:
>
>     np.issubdtype(a.dtype, float) or np.issubdtype(a.dtype, int) or
>     np.issubdtype(a.dtype, complex)
>
>     And for string:
>
>     np.issubdtype(a.dtype, str)
>
>
> These are valid but what I don't like is that I need to know the list 
> of possible number types. Basically I don't like a test that fails 
> because I didn't know about a dtype. For string It is ok, the universe 
> of is either string or not string. Maybe this is as good as it gets.
>
> I guess my use case is that I want to be sure I can perform math on 
> the values. So maybe I should just do someting like 
> "numpy.lib._iotools._is_string_like" but "_is_number_like", Maybe 
> there is such and I missed it. If not there should be.
>
> Vincent
>
>
>
Can you give an example of what you are trying to do?

If some of your string arrays only have string representations of 
numbers that you want to do the math on then you have to attempt to 
convert those arrays into a numeric dtype (probably float) using for 
example asarray().

Bruce

 >>> import numpy as np
 >>> a=np.array([1,2,3])
 >>> c=np.array(['1','2','3'])
 >>> d=np.array(['a','b','1'])
 >>> np.asarray(a, dtype=float)
array([ 1.,  2.,  3.])
 >>> np.asarray(c,dtype=float)
array([ 1.,  2.,  3.])
 >>> np.asarray(d,dtype=float)
Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
   File "/usr/lib64/python2.6/site-packages/numpy/core/numeric.py", line 
284, in asarray
     return array(a, dtype, copy=False, order=order)
ValueError: invalid literal for float(): a
 >>> try:
...     np.asarray(d,dtype=float)
... except:
...     print 'fail'
...
fail


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20100527/2a3e811b/attachment.html>