Determines the sparse difference between the subject and reference datasets, by computing subject-reference.
Both datasets must contain parameters x,y, and z. The subject dataset must contain x,y,z values that matches exactly with a section of the x,y,z values of the reference dataset. Each dataset may have one additional parameter that must match exactly.
subject_dataset: The subject Lumerical dataset.
reference_dataset: The reference Lumerical dataset.
out: A struct of structs containing the sparse difference in compressed sparse column (CSC) format for each dataset attribute. Each structure corresponds to a single attribute in the rectilinear dataset. In the CSC format, the column indices (1-based) represent the indices of the additional parameter, and the row indices (1-based) corresponds to a linear index converted from the \(x,y,z\) data in the reference dataset. For example, the \(i,j,k\)-th element in \(x,y,z\) in the reference dataset is converted to a linear row index via \(i+(j-1)\times\text{len}(x)+(k-1)\times\text{len}(x)\times\text{len}(y)\).