[Numpy-discussion] Interleaved Arrays and

[Robert]

Neil Martinsen-Burrell wrote:
> On 06/16/2009 02:18 PM, Robert wrote:
>>   >>>  n = 10
>>   >>>  xx = np.ones(n)
>>   >>>  yy = np.arange(n)
>>   >>>  aa = np.column_stack((xx,yy))
>>   >>>  bb = np.column_stack((xx+1,yy))
>>   >>>  aa
>> array([[ 1.,  0.],
>>          [ 1.,  1.],
>>          [ 1.,  2.],
>>          [ 1.,  3.],
>>          [ 1.,  4.],
>>          [ 1.,  5.],
>>          [ 1.,  6.],
>>          [ 1.,  7.],
>>          [ 1.,  8.],
>>          [ 1.,  9.]])
>>   >>>  bb
>> array([[ 2.,  0.],
>>          [ 2.,  1.],
>>          [ 2.,  2.],
>>          [ 2.,  3.],
>>          [ 2.,  4.],
>>          [ 2.,  5.],
>>          [ 2.,  6.],
>>          [ 2.,  7.],
>>          [ 2.,  8.],
>>          [ 2.,  9.]])
>>   >>>  np.column_stack((aa,bb))
>> array([[ 1.,  0.,  2.,  0.],
>>          [ 1.,  1.,  2.,  1.],
>>          [ 1.,  2.,  2.,  2.],
>>          [ 1.,  3.,  2.,  3.],
>>          [ 1.,  4.,  2.,  4.],
>>          [ 1.,  5.,  2.,  5.],
>>          [ 1.,  6.,  2.,  6.],
>>          [ 1.,  7.,  2.,  7.],
>>          [ 1.,  8.,  2.,  8.],
>>          [ 1.,  9.,  2.,  9.]])
>>   >>>  cc = _
>>   >>>  cc.reshape((n*2,2))
>> array([[ 1.,  0.],
>>          [ 2.,  0.],
>>          [ 1.,  1.],
>>          [ 2.,  1.],
>>          [ 1.,  2.],
>>          [ 2.,  2.],
>>          [ 1.,  3.],
>>          [ 2.,  3.],
>>          [ 1.,  4.],
>>          [ 2.,  4.],
>>          [ 1.,  5.],
>>          [ 2.,  5.],
>>          [ 1.,  6.],
>>          [ 2.,  6.],
>>          [ 1.,  7.],
>>          [ 2.,  7.],
>>          [ 1.,  8.],
>>          [ 2.,  8.],
>>          [ 1.,  9.],
>>          [ 2.,  9.]])
>>   >>>
>>
>>
>> However I feel too, there is a intuitive abbrev function like
>> 'interleave' or so missing in numpy shape_base or so.
> 
> Using fancy indexing, you can set strided portions of an array equal to 
> another array.  So::
> 
> In [2]: aa = np.empty((10,2))
> 
> In [3]: aa[:, 0] = 1
> 
> In [4]: aa[:,1] = np.arange(10)
> 
> In [5]: bb = np.empty((10,2))
> 
> In [6]: bb[:,0] = 2
> 
> In [7]: bb[:,1] = aa[:,1] # this works
> 
> In [8]: cc = np.empty((20,2))
> 
> In [9]: cc[::2,:] = aa
> 
> In [10]: cc[1::2,:] = bb
> 
> In [11]: cc
> Out[11]:
> array([[ 1.,  0.],
>         [ 2.,  0.],
>         [ 1.,  1.],
>         [ 2.,  1.],
>         [ 1.,  2.],
>         [ 2.,  2.],
>         [ 1.,  3.],
>         [ 2.,  3.],
>         [ 1.,  4.],
>         [ 2.,  4.],
>         [ 1.,  5.],
>         [ 2.,  5.],
>         [ 1.,  6.],
>         [ 2.,  6.],
>         [ 1.,  7.],
>         [ 2.,  7.],
>         [ 1.,  8.],
>         [ 2.,  8.],
>         [ 1.,  9.],
>         [ 2.,  9.]])
> 
> Using this syntax, interleave could be a one-liner.
> 
> -Neil

that method of 'filling an empty with a pattern' was mentioned in 
the other (general) interleaving question. It requires however a 
lot of particular numbers and :'s in the code, and requires even 
more statements which can hardly be written in functional style - 
in one line?. The other approach is more jount, free of fancy 
indexing assignments.

The general interleaving should work efficiently in one like this:

np.column_stack/concatenate((r,g,b,....), axis=...).reshape(..)


But as all this is not intuitive, something like this should be in 
numpy perhaps? :

def interleave( tup_arrays, axis = None )


Robert