R/egdt.R

In NMdata: Preparation, Checking and Post-Processing Data for PK/PD Modeling

##' Expand grid of data.tables
##'
##' @param dt1 a data.table.
##' @param dt2 another data.table.
##' @param quiet The default is to give some information along the way
##'     on what data is found. But consider setting this to TRUE for
##'     non-interactive use. Default can be configured using
##'     NMdataConf.
##' @details Merging works mostly similarly for data.table and
##'     data.table. However, for data.table the merge must be done by
##'     one or more columns. This means that the convenient way to
##'     expand all combinations of all rows in two data.frames is not
##'     available for data.tables. This functions provides that
##'     functionality. It always returns data.tables.
##' @return a data.table that expands combinations of rows in dt1 and
##'     dt2.
##' @examples
##' df1 <- data.frame(a=1:2,b=3:4)
##' df2 <- data.frame(c=5:6,d=7:8)
##' merge(df1,df2)
##' library(data.table)
##' ## This is not possible
##' \dontrun{
##' merge(as.data.table(df1),as.data.table(df2),allow.cartesian=TRUE)
##' }
##' ## Use egdt instead
##' egdt(as.data.table(df1),as.data.table(df2))
##' @import data.table

##' @export

egdt <- function(dt1,dt2,quiet){

    if(missing(quiet)) quiet <- NULL
    quiet <- NMdataDecideOption("quiet",quiet)
    
    dt1 <- copy(as.data.table(dt1))
    dt2 <- copy(as.data.table(dt2))
    ## check for common columns
    cols.common <- intersect(colnames(dt1),colnames(dt2))
    if(length(cols.common)>0) messageWrap(sprintf("Common column names in dt1 and dt2. If a merge is intended, it is beyond the scope of egdt. If you still want to use egdt, you can rename the common column names first. Common column names: %s.",paste(cols.common,collapse=", ")),fun.msg=stop)
    tc <- tmpcol(names=c(colnames(dt1),colnames(dt2)))

    dt1[,(tc):=1L]
    dt2[,(tc):=1L]

    result <- merge(dt1,dt2,by=tc,allow.cartesian = TRUE)
    result[,(tc):=NULL]

    if(!quiet){
        dt1[,(tc):=NULL]
        dt2[,(tc):=NULL]    
        print(dims(dt1,dt2,result))
    }
    result[]
}