[SciPy-user] Behavior of string.join differs from standard python

Travis Oliphant oliphant.travis at ieee.org
Tue Jan 24 12:06:20 CST 2006

Yann Le Du wrote:

>Just a small problem I stumbled upon.
>I'd like to know why the behavior of string.join in scipy is different 
>from that of standard python ?
>In scipy :
># string.join("",["a","b","c"])
Look at




for part of the answer.

>And also, the documentation of string.join is not clear, because if you do 
>what's written there, there's an error.
The doc-string (please call it that not "the documentation" which 
usually refers to off-line docs), is exactly the same as str.join 
because the string object in numpy inherits from the str object in Python.

So, numpy.string.join is an unbound method.  Thus, the first argument 
must be a "joining" string.  However, the string module has a different 
calling syntax.

My understanding is that the string module is deprecated, by the way 
since most of the functionality are now methods on strings.


More information about the SciPy-user mailing list