When translating a decision tree from sklearn to Spark ML, an error occurs stating that the maxBins parameter should be at least equal to the number of values in each categorical feature. Why does Spark ML have this requirement? Choose only ONE best answer.
