R/by_sampleBy.R

In doBy: Groupwise Statistics, LSmeans, Linear Estimates, Utilities

#########################################################################
#' @title Sampling from a data frame
#' @description A data frame is split according to some variables in a
#'     formula, and a sample of a certain fraction of each is drawn.
#' @name by-sample
#########################################################################
#' @details If systematic=FALSE (default) then frac gives the fraction
#'     of data sampled. If systematic=TRUE and frac=.2 then every 1/.2
#'     i.e. every 5th observation is taken out.
#'
#' @param formula A formula defining the grouping of the data frame.
#' @param frac The part of data to be sampled.
#' @param replace Is the sampling with replacement.
#' @param data A data frame.
#' @param systematic Should sampling be systematic.
#'
#' @return A dataframe.
#' @seealso \code{\link{orderBy}}, \code{\link{order_by}},
#'    \code{\link{splitBy}}, \code{\link{split_by}}, 
#'    \code{\link{summaryBy}}, \code{\link{summary_by}}, 
#'    \code{\link{transformBy}}, \code{\link{transform_by}}
#'
#' @keywords utilities
#' 
#' @examples
#' data(dietox)
#' sampleBy(formula = ~ Evit + Cu, frac=.1, data = dietox)


#' @export
#' @rdname by-sample
sample_by <- function(data, formula,
                      frac=0.1, replace=FALSE, systematic=FALSE){
    arg <- list(formula=formula, frac=frac, replace=replace, data=data,
                systematic=systematic) 
    do.call(sampleBy, arg)
}


#' @export
#' @rdname by-sample
sampleBy <- function(formula, frac=0.1, replace=FALSE, data=parent.frame(),
                     systematic=FALSE){

    out <- splitBy(formula, data=data)

    out <- lapply(out, function(dat){
        if (systematic){
            idx <- seq(1, nrow(dat), 1 / frac)
        } else {  
            idx <- sort(sample(1:nrow(dat), size=round(frac * nrow(dat)),
                               replace=replace))
        }
        dat[idx, ]
    }) 
    out <- do.call("rbind", out)
    out
}