如何检查Python csv.writer实例的类？

Question

I have a function that will write a python list to various formats. csv is one of the formats I would like to write out to. I know i use the built-in csv Python module, but my function is currently written in the following manner:

def foo(self, save_to=None):
    if save_to is None:
        print self.my_list

I can check the type of the csv.writer object by doing:

writer = csv.writer(open('test.csv','w'))
writer
<_csv.writer at 0x7f278125aca8>    

type(writer)
_csv.writer

How do I check to see in my function if save_to is an csv object? Usually I'd use the isinstance in a function function like such:

def foo(self, save_to=None):
    if save_to is None:
        print self.my_list
    if isinstance(save_to, _csv.writer):
        save_to.writerow([self.my_list])

but I'm getting a NameError and AttrubuteError

In [28]: isinstance(writer, _csv.writer)
---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-28-0af1f54bd092> in <module>()
----> 1 isinstance(writer, _csv.writer)    

NameError: name '_csv' is not defined    

In [29]: isinstance(writer, csv._csv.writer)
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-29-4e8e62254caf> in <module>()
----> 1 isinstance(writer, csv._csv.writer)    

AttributeError: 'module' object has no attribute '_csv

Answer 1

You shouldn't use an isinstance check here (and especially not with a class in a private module like _csv ): isinstance calls violate OOP in general. More specific to Python, its "duck-typing" means that you should just try and use the method and catch attribute errors: if it quacks like a duck, it's a duck -- if it implements writerow like a csv writer, it's a csv writer.

try:
    save_to.writerow([self.my_list])
except AttributeError:
    # not a csv writer, do something else

More specifically though, the reason you're getting the errors you detail is because the submodule isn't imported.