Preprocess() function with full sample (sample = 1.0)

[OttoEctrie99]

Preprocess() function with full sample (sample = 1.0) does not result in train and test samples sizes which are jointly equivalent to full sample size of the data given as input. More specifically, I pass a data set with 3016 observations of 150 variables to the preprocess() function, but the resulting data list only contains a train and test set of 970 and 415 observations respectively. The same holds for the Adulte example. Is this due to the balancing incorporated in the function code?