df = df.sample(frac=1).reset_index(drop=True) #Use if you want to reset index order df.sample(frac=1) # Use for no reset in index order
shuffled_train_df = train_df.reindex(np.random.permutation(train_df.index))