CODEWITHSUNDEEP
pythonpython-3.xnumpynumpy-indexing
E. Sommer
E. Sommer

Reputation: 750

Numpy advanced indexing fails

I have a numpy array looking like this:

 a = np.array([[0.87, 1.10, 2.01, 0.81 , 0.64,        0.        ],
   [0.87, 1.10, 2.01, 0.81 , 0.64,        0.        ],
   [0.87, 1.10, 2.01, 0.81 , 0.64,        0.        ],
   [0.87, 1.10, 2.01, 0.81 , 0.64,        0.        ],
   [0.87, 1.10, 2.01, 0.81 , 0.64,        0.        ],
   [0.87, 1.10, 2.01, 0.81 , 0.64,        0.        ]])

I like to manipulate this by setting the 'bottom left' part to zero. Instead of looping through rows and columns, I want to achieve this by means of indexing: 

ix = np.array([[1, 1, 1, 1, 1, 1],
   [0, 1, 1, 1, 1, 1],
   [0, 0, 1, 1, 1, 1],
   [0, 0, 0, 1, 1, 1],
   [0, 0, 0, 0, 1, 1],
   [0, 0, 0, 0, 0, 1]])

However a[ix] does not deliver what I expect, as a[ix].shape is now (6,6,6), i.e. a new dimension has been added. What do I need to do in order to preserve the shape of a, but with all zeros in the bottom left?

Upvotes: 1

Views: 71

Answers (2)

user3483203
user3483203

Reputation: 51155

If you don't want to have to worry about creating ix at all, what you're really asking for is the upper triangle of a, which is the method numpy.triu

np.triu(a)


array([[0.87, 1.1 , 2.01, 0.81, 0.64, 0.  ],
       [0.  , 1.1 , 2.01, 0.81, 0.64, 0.  ],
       [0.  , 0.  , 2.01, 0.81, 0.64, 0.  ],
       [0.  , 0.  , 0.  , 0.81, 0.64, 0.  ],
       [0.  , 0.  , 0.  , 0.  , 0.64, 0.  ],
       [0.  , 0.  , 0.  , 0.  , 0.  , 0.  ]])

Upvotes: 1

akuiper
akuiper

Reputation: 214967

You don't need advanced indexing for this purpose. Boolean indexing would be more suitable with what you have:

a[~ix.astype(bool)] = 0
a
#array([[ 0.87,  1.1 ,  2.01,  0.81,  0.64,  0.  ],
#       [ 0.  ,  1.1 ,  2.01,  0.81,  0.64,  0.  ],
#       [ 0.  ,  0.  ,  2.01,  0.81,  0.64,  0.  ],
#       [ 0.  ,  0.  ,  0.  ,  0.81,  0.64,  0.  ],
#       [ 0.  ,  0.  ,  0.  ,  0.  ,  0.64,  0.  ],
#       [ 0.  ,  0.  ,  0.  ,  0.  ,  0.  ,  0.  ]])

Upvotes: 1

Related Questions