R/counting_events_group.R

In CorReg-community/CorReg: Linear Regression Based on Linear Structure Between Variables

#'Counts events
#' 
#' @description creates a data.frame counting occurencies of specific events
#' in a given vector separated by named connex groups.
#' For example it allows to count occurencies of each category of a qualitative
#'  variable above in the dataset for each individual. 
#'  Grouping allows to reset the count each time another variable changes.
#'  
#' The idea is to create a cumulative history of what happened before 
#' current individual.
#' 
#' One new variable is created for each unique value of events
#' 
#' @param events The vector of events to count (qualitative values). 
#' One new variable is created for each unique value of events.
#' @param group qualitative vector that resets counting when its value changes.
#' @param base a dataframe to which counting can be added (cbind).
#' @param prefix (optional) prefix to the newly created variables names
#' @param suffix (optional) suffix to the newly created variables names
#' @param clean (boolean) to decide either or not variables names has to be 
#' cleaned (spaces, accents and special characters)
#' @param keep (boolean) to decide either or not events and group are to 
#' be added to the resulting data.frame
#' @return a data.frame which number of rows is the length of events
#' 
#' @examples 
#' events <- rep(c('a', 'b', 'c'), times = 20)
#' group <- rep(rep(c('A', 'B'), c(12, 8)), times = 3)
#' counting_events_group(events, group)
#' 
#' @export
counting_events_group <- function(events = events, group = NULL, base = NULL, prefix = NULL, suffix = NULL, clean = TRUE, keep = FALSE) {
    # base=base events=base$cat_aci group=base$NO_PRG table(base$cat_aci,base$fam_pro_vis)
    if (is.null(group)) {
        group = rep(1, times = length(events))
    }
    families = unique(events)
    ## Initialisation
    additional = data.frame(cbind(group, events))  #creates a data.frame same nb of rows
    
    # for each family :
    for (family in families) {
        additional = cbind(additional, 0)
        # create a vector (list element or matrix,...)
    }
    # rename them if necessary (using families list)
    names(additional)[-c(1, 2)] = paste0(prefix, as.character(families), suffix)
    if (clean == TRUE) {
        names(additional) = clean_symbols(names(additional))
    }
    
    # define a variable to store local Montage (inital value = first met after timestamp ordering)
    local_prog = group[1]
    
    ## Core code For each coil
    for (i in 2:nrow(additional)) {
        # if new montage=> all values to 0
        if (group[i] != group[i - 1]) {
            # set local Montage value to actual
            local_prog = group[i]
        } else {
            # else (same montage) => copy previous line
            additional[i, -c(1, 2)] = additional[i - 1, -c(1, 2)]
            # add one to the element corresponding to previous value
            family_loc = which(families == events[i - 1])  #get previous family
            additional[i, family_loc + 2] = additional[i, family_loc + 2] + 1
        }
    }
    if (keep == FALSE) {
        # don't keep columns used for counting
        additional = additional[, -c(1, 2)]
    }
    # add new elements to the dataset
    if (!is.null(base)) {
        additional = cbind(base, additional)
    }
    return(additional)
}