python array indexing list in list

Question

I want to do array indexing. I would have expected the result to be [0,1,1,0], however I just get an error. How can I do this type of indexing?

a_np_array=np.array(['a','b','c','d'])
print a_np_array in ['b', 'c']

Traceback (most recent call last):
File "dfutmgmt_alpha_osis.py", line 130, in <module>
print a_np_array in ['b', 'c']
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

Up top, i actually meant to say [False,True,True,False] not [0,1,1,0] as I want the bools so i can do indexing

OK. I checked the documentation and this seems to be what I am looking for. Can you explain why my method does not work? I am surprised that there is no kind of array index like method to do this. — user2015487
– user2015487, Commented Mar 2, 2015 at 16:27
I have moved my comment to answer with some more explanation. — Ashwini Chaudhary
– Ashwini Chaudhary, Commented Mar 2, 2015 at 17:32

Reut Sharabani · Accepted Answer · 2015-03-02 16:22:07Z

2

Try this list comprehension:

>>> print [int(x in ['b', 'c']) for x in a_np_array]
[0, 1, 1, 0]

Utilizing the fact that int(True) == 1 and int(False) == 0

answered Mar 2, 2015 at 16:22

Reut Sharabani

31.5k7 gold badges76 silver badges95 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Ashwini Chaudhary · Accepted Answer · 2015-03-02 17:45:07Z

First of all you cannot use [0,1,1,0] for indexing here, so you're using the wrong term.

>>> a_np_array[[0,1,1,0]]  # Not useful at all
array(['a', 'b', 'b', 'a'], 
      dtype='|S1')

If I understood this correctly you're simply trying to check whether items of a_np_array exist in ['b', 'c'], for that use numpy.in1d, but as it returns boolean array we just need to convert it integer array.

>>> np.in1d(a_np_array, ['b','c'])
array([False,  True,  True, False], dtype=bool)
>>> np.in1d(a_np_array, ['b','c']).astype(int)
array([0, 1, 1, 0])

Coming to why a_np_array in ['b', 'c'] didn't work?

Here the in operator will call the __contains__ method of the list object(['b', 'c']) and then for each object in list Python will use the method PyObject_RichCompareBool to compare each item to a_np_array. PyObject_RichCompareBool first of all simply checks if the items to be compared are the same object, i.e same id(), if yes return 1 right away otherwise call PyObject_RichCompare on them. Hence this will work:

>>> a_np_array in [a_np_array, 'c', 'a']
True

But this won't:

>>> a_np_array in [a_np_array.copy(), 'c', 'a']
Traceback (most recent call last):
  File "<ipython-input-405-dfe2729bd10b>", line 1, in <module>
    a_np_array in [a_np_array.copy(), 'c', 'a']
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

Now Python will check whether the object returned by PyObject_RichCompare is already a boolean type, i.e True or False(This is done using PyBool_Check and Py_True), if it is then return the result immediately otherwise call PyObject_IsTrue to check whether the object can be considered a truthy object, this is done by calling __nonzero__ method of the object. For a NumPy array this will end up calling bool() on the returned object which is going to raise an error you're getting. Here NumPy expects you to call either all() or any() to check whether all items are True or at least one.

>>> bool(a_np_array == 'a')
Traceback (most recent call last):
  File "<ipython-input-403-b7ced85c4f02>", line 1, in <module>
    bool(a_np_array == 'a')
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

Links to source code:

Collectives™ on Stack Overflow

python array indexing list in list

2 Answers 2

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related