Check if variable is defined with Numpy array?

Question

Sometimes I have a situation where I want to test whether a variable is 0 or None or not. In pure Python, this is simply

foo == True

but when foo is possibly a Numpy object (such as numpy.ndarray), this does not work anymore and I get the error:

ValueError: The truth value of an array with more than one element is ambiguous. 
 Use a.any() or a.all()

and in this case I want a.any(), however this fails on non-iterable objects.

Of course I can explicitly check if foo is 0 or None, but I am wondering if there is a more elegant way to do this.

How can I check if an arbitrary object (both iterable and not) is True or not?

More Pythonically, it's `if foo` - and that'll work with most things. Unfortunately, you'll have to special case `numpy.ndarray` or any other type that's more complex than 0, None, False, len() == 0 equates to False. — Jon Clements, Oct 19 '12 at 12:08
I usually check numpy object to be None with `array is None`, but this doesn't check for zero though. See here for a similar question, http://stackoverflow.com/questions/5086178/python-how-to-check-if-array-is-not-empty — gg349, Oct 19 '12 at 14:05

score 2 · Accepted Answer · edited Oct 24 '12 at 19:17

2

Just use np.all or np.any instead of the .all method. np.all will create an array if it is not one yet.

edited Oct 24 '12 at 19:17

the Tin Man

answered Oct 19 '12 at 12:09

seberg

1

Just be aware that this doesn't work as you might think for all iterables: `np.any(False for _ in range(10))` yields `True`. (Of course, it would work if this were a list comprehension instead of a generator.) Basically, it evaluates the logical truth of the iterable instance instead of making an array from the contents. You can work around this with `np.fromiter`, but it's an extra step. Still, using `np.all`, etc, is by far the cleanest solution if you just want to deal with most cases. – Joe Kington Oct 19 '12 at 14:51
This seems like a suitable solution, thanks! The generator exception is not really an issue for me, but it's good to know there are potential pitfalls. – Tim Oct 24 '12 at 19:12

score 1 · Answer 2 · answered Oct 19 '12 at 12:07

1

The recipe I use is to surround the statement assuming the existence of a variable in a try except block:

try:
    do_something(foo)
except NameError:
    foo = something

answered Oct 19 '12 at 12:07

Ben Allison

2 Answers2