Exception if key is a tuple or other value with ns.IC about natsort HOT 6 CLOSED

Can you post the full traceback?

from natsort.

C:\Windows\System32>python
Python 2.7.5 (default, May 15 2013, 22:43:36) [MSC v.1500 32 bit (Intel)] on win
32
Type "help", "copyright", "credits" or "license" for more information.

test = [('hi','1'), ('appe','2'), ('apple',3)]
from natsort import natsort_keygen, ns
test.sort(key = natsort_keygen(key=lambda e: (e[0],e[1]), alg=ns.IC))
Traceback (most recent call last):
File "", line 1, in
File "C:\Python275\lib\site-packages\natsort\natsort.py", line 390, in _natsor
t_key
val = lowercase(val)
File "C:\Python275\lib\site-packages\natsort\locale_help.py", line 79, in
lowercase = lambda x: x.lower()
AttributeError: 'tuple' object has no attribute 'lower'

Hope that example helps! I believe this wont only happen with tuples but with any other data type that is not a string.

from natsort.

This code is in a try: except: block that catches a TypeError. The idea is that the code would assume a string was given, and if not, go to the branch that handles tuples/lists. Clearly, I need to update this to catch an AttributeError as well since calling lower() on a tuple does not raise a TypeError.

from natsort.

Glad I could help! Your library is very useful :)

May I take this chance to ask a question, what is the performance
improvement when using fastnumbers library? I ask this because my proyect
is hosted on amazon elastic beanstalk which is a bit picky about installing
libraries, and fastnumbers needs some C code as well which would be
difficult to install on the cloud. So I was wondering if the performance
improvement is worth it or not.

On Thu, Sep 25, 2014 at 12:26 PM, Seth Morton [email protected]
wrote:

This code is in a try: except: block that catches a TypeError. The idea
is that the code would assume a string was given, and if not, go to the
branch that handles tuples/lists. Clearly, I need to update this to catch
an AttributeError as well.

—
Reply to this email directly or view it on GitHub
#15 (comment).

Cristiano Coelho

_DEVELOPER_Allied South American Professionals
8 de octubre 2323 Of. 1205 - CP 11200
Tel. (+598) 2 408 1699
[email protected] [email protected]
www.asapuruguay.com

from natsort.

You should see about 30% speed improvement in the case where your strings contain only numbers; put another way, it will increase the efficiency of the string to number conversion only. Of course, in real life you are doing other string manipulations that will cause slowdowns that shadow the speed improvements of the number conversion, so the true speed improvement is really dependent on the input data and the options you choose. I recommend installing locally and testing out with your actual data.

The C code shouldn't be a problem to install... it has no library dependencies (except for Python), so if your server has a C compiler that should be all you need.

from natsort.

Check out natsort version 3.5.1, which includes this bugfix.

from natsort.