Why Column Vectors Should be Returned as Matrices in MATLAB

When working with matrix operations in MATLAB, it can sometimes be confusing to determine whether a vector should be treated as a row or column. In some cases, returning a matrix with unitary dimensions may make more sense than dropping these dimensions from method returns. This is particularly true when working with high-dimensional arrays.

In MATLAB, iterating through the columns of a list can be useful if you don't need to know which element you're currently working on. For example, if you have a nx1 vector and want to transpose it, you may need to do so before proceeding. This can be particularly important when working with large datasets where knowing which element you're working on is not crucial.

Inconsistencies between row and column major can also lead to confusion, as one of my laboratory colleagues recently discovered while using Julia to make a small program. He found that the inconsistency in matrix orientation caused him to lose sanity and waste development time.

To illustrate this point, consider the following example. I have a Matlab matrix `data` with dimensions (22,5,82), where columns 1:4 are filled with data for every level of dim(2) and dim(3). When using the `eachrow` function over an array, it does not return rows as expected. Instead, it returns the entire array.

In this case, returning a matrix with unitary dimensions would have been more intuitive and easier to work with. I plan to work on enhancing this functionality in the coming weeks to make it more user-friendly.

Overall, while MATLAB can be powerful for certain tasks, inconsistencies between row and column major can lead to confusion and wasted time. By returning column vectors as matrices, we can simplify matrix operations and make our code more readable and maintainable.

Column vector to matrix matlab