R/build_work_hoursWorkedAllJobs_pnad.R

In antrologos/harmonizePNAD: harmonized variables for the National Household Sample Survey (PNAD-IBGE, Brazil)

build_work_hoursWorkedAllJobs_pnad <- function(Data){

        just_created_vars_list_existedBefore <- exists(x = "just_created_vars", where = .GlobalEnv)

        # Loading the crosswalk
        file_location <- system.file("extdata",
                                     "crosswalk_pnad_hoursWorked.csv",
                                     package = "harmonizePNAD")
        crosswalk   <- data.table::fread(file_location)

        # Selecting the appropriate crosswalk for the current year
        metadata    <- harmonizePNAD:::get_metadata(Data)
        crosswalk_i <- crosswalk[year == metadata$year]

        if(metadata$year %in% 1976:1990){
                # Checking the variable availability
                harmonizePNAD:::check_necessary_vars(Data, crosswalk_i$var_hoursWorkedAllJobs)

                # Recoding
                Data$hoursWorkedAllJobs <- Data[[crosswalk_i$var_hoursWorkedAllJobs]]
                Data[hoursWorkedAllJobs == 999, hoursWorkedAllJobs := NA]
        }

        if(metadata$year %in% 1976:1988){
                Data[hoursWorkedAllJobs == 99, hoursWorkedAllJobs := NA]
        }


        # Correcting hours worked in all jobs
        if(metadata$year %in% 1981:1990){
                harmonizePNAD:::check_necessary_vars(Data, crosswalk_i$var_hoursWorkedMainJob)
                Data$hoursWorkedMainJob_tmp <- Data[[crosswalk_i$var_hoursWorkedMainJob]]

                Data[hoursWorkedAllJobs == 0, hoursWorkedAllJobs := hoursWorkedMainJob_tmp]
                Data[, hoursWorkedMainJob_tmp := NULL]
        }

        if(metadata$year %in% 1992:2015){

                vars <- with(crosswalk_i, c(var_hoursWorkedMainJob,
                                            var_hoursWorkedSecondJob,
                                            var_hoursWorkedThirdJob))

                # Checking the variable availability
                harmonizePNAD:::check_necessary_vars(Data, vars)

                # Recoding
                matrix_hours <- as.matrix(Data[, vars, with = F])

                all_na       <- apply(matrix_hours, 1, function(x) sum(is.na(x)) == ncol(matrix_hours))

                missing_location = NULL
                for(j in 1:ncol(matrix_hours)){
                        missing_location_j <- which(matrix_hours[,j] == 99)

                        if(length(missing_location_j) > 0){
                                missing_location <- rbind(missing_location,
                                                          cbind(missing_location_j, j)
                                )
                        }
                }

                if(!is.null(missing_location)){
                        matrix_hours[missing_location] <- NA
                }

                Data[, hoursWorkedAllJobs := rowSums(matrix_hours, na.rm = T)]
                Data[all_na, hoursWorkedAllJobs := NA]

                if(!is.null(missing_location)){
                        Data[missing_location[,1] , hoursWorkedAllJobs := ifelse(hoursWorkedAllJobs == 0, NA, hoursWorkedAllJobs)]
                }
        }

        Data <- harmonizePNAD:::check_and_build_onTheFly(Data,
                                                         var_name = "occupationalStatus",
                                                         general_or_specific = "general")

        Data[is.na(occupationalStatus) | occupationalStatus == 0, hoursWorkedAllJobs := NA]

        # Truncating
        Data[hoursWorkedAllJobs > 98, hoursWorkedAllJobs := 98]

        if(just_created_vars_list_existedBefore == F){
                Data <- harmonizePNAD:::erase_just_created_vars(Data)
        }


        gc()

        Data
}