如何从具有矩阵值的字典中提取列数据？

Question

I have a dictionary where the values are 3 column matrices. I want to extract the third column of each value and concatenate them all together side by side into a single matrix. I can't figure out how to do this as general slicing won't work on a dictionary. Would dictionaries even be the best way to store this data or would Pandas be better? I've tried doing list(my_dict.values()) but this doesn't seem to work anymore. Running python 3.6.

Answer 1

In [1]: dd = {'one':np.arange(12).reshape(4,3),'two':np.arange(12).reshape(4,3)*2}

If key order doesn't matter:

In [3]: dd.values()
Out[3]: 
dict_values([array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]]), array([[ 0,  2,  4],
       [ 6,  8, 10],
       [12, 14, 16],
       [18, 20, 22]])])

Not what we want (in PY3)

In [4]: np.array(dd.values())
Out[4]: 
array(dict_values([array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]]), array([[ 0,  2,  4],
       [ 6,  8, 10],
       [12, 14, 16],
       [18, 20, 22]])]), dtype=object)

list(values()) looks good - giving a list of arrays which can be joined in different ways

In [5]: np.array(list(dd.values()))
Out[5]: 
array([[[ 0,  1,  2],
        [ 3,  4,  5],
        [ 6,  7,  8],
        [ 9, 10, 11]],

       [[ 0,  2,  4],
        [ 6,  8, 10],
        [12, 14, 16],
        [18, 20, 22]]])

In [7]: np.array(list(dd.values()))[:,:,2]
Out[7]: 
array([[ 2,  5,  8, 11],
       [ 4, 10, 16, 22]])

np.concatenate(list(dd.values()),axis=0) makes a 2d array instead of 3d, so the [:,2] would be 1d instead of 2d. But other wise the same.