The K-S test assumes the data comes from a continuous distribution, so count data would mess up the test's false positive rate. You could avoid that by doing a

The K-S test assumes the data comes from a continuous distribution, so count data would mess up the test's false positive rate. You could avoid that by doing a permutation test, or one of the variations of the K-S test designed to account for data with ties.