Create a non-uniform 2-D vector in python

Question

My objective is to create the following table (2D vector of namedtuples) in python:

     (0,0,1) | (2,0,7) | (1,0,1)
             | (3,0,1) | (4,0,1)
             | (5,0,1) |

As can be seen, table T[0] is of size 1, T[1] is of size 3 and T[2] of size 2. All approches I have seen so far relate to the use List Comprehension but that results in a fixed MxN matrix. I am curious to know if I can use python to first pre-allocate the size of first-dimension (in my case 3) and then keep appending elements to each T[i].

While writing in C++, I would declared a vector T of size 3 and continue to push_back elements to each dimension as need be in a dynamic fashion. However I am unable to find an elegant way to write this is python.

I am very new to python and my first attempt was to do the following (without the use of namedtuples):

 T=[]
 arr1=[0,1,2]
 arr2=[3,4]
 arr3=[6,6,7,8]

 T.append(arr1)
 T.append(arr2)
 T.append(arr3)

This led to the following result:

  >>T
  [[0, 1, 2], [3, 4], [5, 6, 7, 8]]

However if I clear arr1, it shall be removed from T as well. Is there a better approach to this problem?

Answer 1

When you append arr1 to T, you are not making a new copy of the arr1 list. You need to make a copy of the list before you append it to T so that when you modify arr1, the list that you added to T is not modified.

To do this, use T.append(list(arr1)). That will make a new copy of arr1 before adding it to T.