cell2mat too slow

Hi community,

I am trying to extract all poses from a struct by:

arrayfun(@(node)(node.pose), nodes, 'UniformOutput', false)

It gives me something like this:

ans =

  50×1 cell array

    {[ -0.2232 6.2129 0.8631 -0.5050 0]}
    {[-0.5066 5.4770 -0.9729 -0.2310 0]}
    {[ -0.2650 6.0300 0.7742 -0.6329 0]}
    {[ -0.0436 5.4497 0.7313 -0.6820 0]}
    {[ -0.0892 5.5517 -0.4583 0.8888 0]}
    {[ 0.4760 6.0669 -0.3153 -0.9490 0]}
    {[ -0.5377 6.2957 0.9279 -0.3729 0]}
    {[-0.6971 5.9071 -0.7730 -0.6344 0]}
    ...

But I want it to be a matrix like this:

ans =

   -0.2232    6.2129    0.8631   -0.5050         0
   -0.5066    5.4770   -0.9729   -0.2310         0
   -0.2650    6.0300    0.7742   -0.6329         0
   -0.0436    5.4497    0.7313   -0.6820         0
   -0.0892    5.5517   -0.4583    0.8888         0
    0.4760    6.0669   -0.3153   -0.9490         0
   -0.5377    6.2957    0.9279   -0.3729         0
   -0.6971    5.9071   -0.7730   -0.6344         0
   ...

So I use cell2mat:

cell2mat(arrayfun(@(node)(node.pose), nodes, 'UniformOutput', false))

But the function cell2mat is spending around 50% of runtime for my program, which slow down quite a lot.

Is there anyway to avoid cell2mat that has an efficient runtime?

Thanks

Answer 1

[Bruno Luong]

Try

cat(1, nodes.pose)

[Stephen23]

https://www.mathworks.com/help/matlab/matlab_prog/comma-separated-lists.html

https://www.mathworks.com/matlabcentral/answers/320713-how-to-operate-on-comma-separated-lists

[Douglas Chiang]

Thanks, this is much faster than cell2mat()

Answer 2

[Yazan]

To speed up the computations, you simply need to avoid using structure arrays.