Considering only 2-D arrays i.e. matrices only, is there a difference between flip(a) and fliplud(a)?

Considering only 2-D arrays i.e. matrices only, is there a difference between flip(a) and flipud(a)?

Answer 1

[Star Strider]

As I read the documentation: No.

Answer 2

[Walter Roberson]

flipud always operates on the first dimension even if it is singular. flip operates on the first non-singular dimension by default.

For example flip(1:5) is [5 4 3 2 1] because the first non-singular dimension is the second dimension, but flipud(1:5) would be [1 2 3 4 5] because the flip is done on the first dimension.

[Seetha Rama Raju Sanapala]

We are talking about matrices. I think if they are matrices, and not vectors, there is no difference.

[Walter Roberson]

In MATLAB, vectors are matrices which happen to be singular (length 1) on all except 1 dimension.

If you are certain that your array has a non-singular first dimension, then flip(X) and flip(X,1) and flipud(X) would be equivalent. However, either flip(X,1) or flipud(X) would be preferred coding as it would indicate to the people reading the code that you specifically considered the possibility that your input might be singular on the first dimension and you definitely want the first dimension. It saves the reader from having to trace the code backwards to prove that it is impossible for the input to have a singular first dimension, which might not be possible as it might rely upon the assumption that an input file has multiple rows.

It is best practice to code the dimension of the operation unless you want it to be flexible about which dimension to use, unless perhaps it is completely obvious from the immediate context which dimension will be used.