In addition, you would not do PCA if you have only one measurement, in this case "returns". PCA doesn't sound like it is headed in a productive direction. So nothing you have said so far fits principal components.
The variables in principal components are multiple measurements (variables) of the objects you are studying (in this case, it sounds like the objects are companies) but yet you arrange the data so that the different companies are now variables, not objects. Principal Components does not work with categorical variables. You seem to be choosing Principal Components Analysis without understanding that this method does not fit the problem.
I would like to be able to analyse the principal components according to the industry that the underlying stocks are in