Question

At the moment my code is as follows

A = [matrix_x[i][:n] for i in xrange(0, n)]
B = [matrix_x[i][n:] for i in xrange(0, n)]
C = [matrix_x[i+n][:n] for i in xrange(0, n)]
D = [matrix_x[i+n][n:] for i in xrange(0, n)]

Is there a better way of doing this, since I am continually looping over the same xrange. In this instance, would if be better to not use a list comprehension and just append the values to each list while in a single for loop.

A,B,C,D = [],[],[],[]
for i in xrange(0,n):
    A.append(matrix_x[i][:n])
    B.append(matrix_x[i][n:])
    ... etc

Second way seems more efficient to me. What way would be more ‘pythonic’ or is there another way I haven’t thought of

Answer 1

I generally prefer to iterate over the list of items itself, rather than over xrange(len(list_of_items)) and working with the i’th item at a time. Here’s how to use zip to look at each (this,next) pair in a sequence, and then build up your lists:

A,B,C,D = [],[],[],[]
for this_,next_ in zip(matrix_x,matrix_x[1:]):
    A.append(this_[:n])
    B.append(this_[n:])
    C.append(next_[:n])
    D.append(next_[n:])

And yes, you can really compact this down to a zip of a zip:

A,B,C,D = zip(*((this_[:n],this_[n:],next_[:n],next_[n:])
                  for this_,next_ in zip(matrix_x,matrix_x[1:])))