[Numpy-discussion] Inconsistent error messages.

[Stéfan van der Walt]

Hi Chris and Charles

2009/5/23 Charles R Harris <charlesr.harris at gmail.com>:
>> Robert thought that should be the default, but I think that means
>> everyone would be forced to check how many items they got every time
>> they read, which is too much code and likely to be forgotten and lead to
>> errors. So I think that an exception should be raised if you ask for n
>> and get less than n, but that there should be a flag that says something
>> like "max_items=n", that indicates that you'll take what you get.
>>
>> I don't like a warning -- it's unconventional to catch those like
>> exceptions -- if you want n items, and you haven't written code to
>> handle less than n, you should get an exception. If you have written
>> code to handle that, that you can use the flag.
>
> I don't like the idea of a warning here either. How about adding a keyword
> 'strict' so that strict=1 means an error is raised if the count isn't
> reached, and strict=0 means any count is acceptable?

The reason I much prefer a warning is that you always get data back,
whether things went wrong or not.  If you throw an error, then you
can't get hold of the last read blocks at all.

I guess a strict flag is OK, but why, if you've got a warning in
place?  Warnings are easy to catch (and this can be documented in
fromfile's docstring):

warnings.simplefilter('error', np.lib.IOWarning)

In Python 2.6 you can use "catch_warnings":

with warnings.catch_warnings(True) as w:
    np.fromfile(...)
    if w: print "Error trying to load file"

Regards
Stéfan