Array indexing in Numpy python

Question

How can I define _SOME CODE_ in the next code fragment in order to get the results shown below?

vector = numpy.array([a,b,c,d])
for i in xrange(4):
 print vector[_SOME CODE_ using i]

It sould give me those results:

[a,b,c]
[a,c,d]
[a,b,d]
[b,c,d]

The order is not important.

• Sven Marnach: What's the intended pattern? Add 1 to some vector components?
• Charles: vector doesn't contain 3, so you cannot get the desired results with indexing alone.

Answer 1

Answer for the edited question:

>>> vector = numpy.array([0, 1, 2, 3])
>>> for i in xrange(4):
... print numpy.r_[vector[:i], vector[i+1:]]
... 
[1 2 3]
[0 2 3]
[0 1 3]
[0 1 2]

Answer for the original question: Here's some random code producing the desired output:

>>> import numpy
>>> vector = numpy.array([0,1,2])
>>> for i in xrange(4):
... print vector + (vector >= i)
... 
[1 2 3]
[0 2 3]
[0 1 3]
[0 1 2]

I've got no idea if this is what you want -- the requirement specification left some room for interpretation.

Answer 2

I think you're trying to find all combinations of size 3 of the vector. You can do this with itertools.combinations like this:

>>> import numpy
>>> import itertools
>>> vector = numpy.array([0, 1, 2, 3])
>>> list(itertools.combinations(vector, 3))
[(0, 1, 2), (0, 1, 3), (0, 2, 3), (1, 2, 3)]