Bayesian variable selection for high‐dimensional rank data