Choosing axis in numpy array using select method

Question

I have following arrays.

In [1]: a = np.array([["aa", "bb", "cc"], ["cc", "bb", "aa"]])
In [2]: a
Out[2]:
array([['aa', 'bb', 'cc'],
       ['cc', 'bb', 'aa']],
       dtype='|S2')

In [3]: b = np.array([[11, 12, 13], [21, 22, 23]])
In [4]: b
Out[4]:
array([[11, 12, 13],
       [21, 22, 23]])

Relation between a and b can be described conceptually as b[0].aa = 11 b[0].bb = 12. b[0].cc = 13. i.e. row in 'a' are keys and row in 'b' are values of a single dict. First rows will represent

{'aa': 11, 'bb': 12, 'cc': 13}

Now we are given with keys of these dicts.

In [5]: c = np.array(["bb", "aa"])
In [6]: c
Out[6]:
array(['bb', 'aa'],
  dtype='|S2')

Now, what is the best way of accessing b array given 'c' which will give axis of each row as value in 'a'. One way to do is

In [7]: cond_list = [a[:, 0] == c,  a[:, 1] == c, a[:, 2] == c]
In [8]: choice_list = [b[:, 0],  b[:, 1], b[:, 2]]
In [9]: np.select(cond_list, choice_list)
Out[9]: array([12, 23])

Is there a better way of doing it? What if number of axis in a and b are not known before hand?

unutbu · Accepted Answer · 2011-08-02 11:53:14Z

2

In [13]: a==c[:,np.newaxis]
Out[13]: 
array([[False,  True, False],
       [False, False,  True]], dtype=bool)

In [14]: b[a==c[:,np.newaxis]]
Out[14]: array([12, 23])

answered Aug 2, 2011 at 11:53

unutbu

886k197 gold badges1.9k silver badges1.7k bronze badges

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

Collectives™ on Stack Overflow

Choosing axis in numpy array using select method

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related