SeriesAggreg <- function(TabSeries,
TimeFormat, NewTimeFormat,
ConvertFun,
YearFirstMonth = 1, TimeLag = 0,
verbose = TRUE) {
##_Arguments_check
##check_TabSeries
if (is.null(TabSeries) ) {
stop("'TabSeries' must be a data.frame containing the dates and data to be converted")
}
if (!is.data.frame(TabSeries)) {
stop("'TabSeries' must be a data.frame containing the dates and data to be converted")
}
if (ncol(TabSeries) < 2) {
stop("'TabSeries' must contain at least two columns (including the coulmn of dates")
}
##check_TimeFormat
if (!any(class(TabSeries[, 1]) %in% "POSIXt")) {
stop("'TabSeries' first column must be a vector of class 'POSIXlt' or 'POSIXct'")
}
if (any(class(TabSeries[, 1]) %in% "POSIXlt")) {
TabSeries[, 1] <- as.POSIXct(TabSeries[, 1])
}
for (iCol in 2:ncol(TabSeries)) {
if (!is.numeric(TabSeries[,iCol])) {
stop("'TabSeries' columns (other than the first one) be of numeric class")
}
}
if (is.null(TimeFormat)) {
stop("'TimeFormat' must be one of 'hourly', 'daily', 'monthly' or 'yearly'")
}
if (!is.vector(TimeFormat)) {
stop("'TimeFormat' must be one of 'hourly', 'daily', 'monthly' or 'yearly'")
}
if (!is.character(TimeFormat)) {
stop("'TimeFormat' must be one of 'hourly', 'daily', 'monthly' or 'yearly'")
}
if (length(TimeFormat) != 1) {
stop("'TimeFormat' must be one of 'hourly', 'daily', 'monthly' or 'yearly'")
}
if (! TimeFormat %in% c("hourly", "daily", "monthly", "yearly")) {
stop("'TimeFormat' must be one of 'hourly', 'daily', 'monthly' or 'yearly'")
}
##check_NewTimeFormat
if (is.null(NewTimeFormat)) {
stop("'NewTimeFormat' must be one of 'hourly', 'daily', 'monthly' or 'yearly'")
}
if (!is.vector(NewTimeFormat)) {
stop("'NewTimeFormat' must be one of 'hourly', 'daily', 'monthly' or 'yearly'")
}
if (!is.character(NewTimeFormat)) {
stop("'NewTimeFormat' must be one of 'hourly', 'daily', 'monthly' or 'yearly'")
}
if (length(NewTimeFormat) != 1) {
stop("'NewTimeFormat' must be one of 'hourly', 'daily', 'monthly' or 'yearly'")
}
if (! NewTimeFormat %in% c("hourly", "daily", "monthly", "yearly")) {
stop("'NewTimeFormat' must be one of 'hourly', 'daily', 'monthly' or 'yearly'")
}
##check_ConvertFun
if (is.null(ConvertFun)) {
stop("'ConvertFun' must be a vector of character")
}
if (!is.vector(ConvertFun)) {
stop("'ConvertFun' must be a vector of character")
}
if (!is.character(ConvertFun)) {
stop("'ConvertFun' must be a vector of character")
}
if (length(ConvertFun) != (ncol(TabSeries) - 1)) {
stop(
paste("'ConvertFun' must be of length", ncol(TabSeries) - 1, "(length=ncol(TabSeries)-1)")
)
}
if (sum(ConvertFun %in% c("sum", "mean") == FALSE) != 0) {
stop("'ConvertFun' elements must be one of 'sum' or 'mean'")
}
##check_YearFirstMonth
if (is.null(YearFirstMonth)) {
stop("'YearFirstMonth' must be an integer between 1 and 12")
}
if (!is.vector(YearFirstMonth)) {
stop("'YearFirstMonth' must be an integer between 1 and 12")
}
if (!is.numeric(YearFirstMonth)) {
stop("'YearFirstMonth' must be an integer between 1 and 12")
}
YearFirstMonth <- as.integer(YearFirstMonth)
if (length(YearFirstMonth) != 1) {
stop("'YearFirstMonth' must be only one integer between 1 and 12")
}
if (YearFirstMonth %in% (1:12) == FALSE) {
stop("'YearFirstMonth' must be only one integer between 1 and 12")
}
##check_DatesR_integrity
if (TimeFormat == "hourly") {
by <- "hours"
}
if (TimeFormat == "daily") {
by <- "days"
}
if (TimeFormat == "monthly") {
by <- "months"
}
if (TimeFormat == "yearly") {
by <- "years"
}
TmpDatesR <- seq(from = TabSeries[1, 1], to = tail(TabSeries[, 1], 1), by = by)
if (!identical(TabSeries[, 1], TmpDatesR)) {
stop("some dates might not be ordered or are missing in 'TabSeries'")
}
##check_conversion_direction
if ((TimeFormat == "daily" & NewTimeFormat %in% c("hourly") ) |
(TimeFormat == "monthly" & NewTimeFormat %in% c("hourly","daily") ) |
(TimeFormat == "yearly" & NewTimeFormat %in% c("hourly","daily","monthly"))) {
stop("only time aggregation can be performed")
}
##check_if_conversion_not_needed
if ((TimeFormat == "hourly" & NewTimeFormat == "hourly" ) |
(TimeFormat == "daily" & NewTimeFormat == "daily" ) |
(TimeFormat == "monthly" & NewTimeFormat == "monthly") |
(TimeFormat == "yearly" & NewTimeFormat == "yearly" )) {
if (verbose) {
warning("the old and new format are identical \n\t -> no time-step conversion was performed")
return(TabSeries)
}
}
##_Time_step_conversion
##_Handle_conventional_difference_between_hourly_series_and_others
TmpDatesR <- TabSeries[, 1]
#if (TimeFormat=="hourly") { TmpDatesR <- TmpDatesR - 60*60; }
TmpDatesR <- TmpDatesR + TimeLag
Hmax <- "00"
if (TimeFormat == "hourly") {
Hmax <- "23"
}
##_Identify_the_part_of_the_series_to_be_aggregated
NDaysInMonth <- list("31", c("28", "29"), "31", "30", "31", "30", "31", "31", "30", "31", "30", "31")
YearLastMonth <- YearFirstMonth + 11
if (YearLastMonth > 12) {
YearLastMonth <- YearLastMonth - 12
}
YearFirstMonthTxt <- formatC(YearFirstMonth, format = "d", width = 2, flag = "0")
YearLastMonthTxt <- formatC(YearLastMonth , format = "d", width = 2, flag = "0")
if (NewTimeFormat == "daily") {
Ind1 <- which(format(TmpDatesR, "%H") == "00")
Ind2 <- which(format(TmpDatesR, "%H") == Hmax)
if (Ind2[1] < Ind1[1]) {
Ind2 <- Ind2[2:length(Ind2)]
}
if (tail(Ind1, 1) > tail(Ind2, 1)) {
Ind1 <- Ind1[1:(length(Ind1) - 1)]
}
### Aggr <- NULL; iii <- 0; for(kkk in 1:length(Ind1)) {
### iii <- iii+1; Aggr <- c(Aggr,rep(iii,length(Ind1[kkk]:Ind2[kkk]))); }
Aggr <- as.numeric(format(TmpDatesR[min(Ind1):max(Ind2)], "%Y%m%d"))
### more efficient
NewDatesR <- data.frame(seq(from = TmpDatesR[min(Ind1)], to = TmpDatesR[max(Ind2)], by = "days"))
}
if (NewTimeFormat=="monthly") {
Ind1 <- which(format(TmpDatesR, "%d%H") == "0100")
Ind2 <- which(format(TmpDatesR,"%m%d%H") %in% paste0(c("0131", "0228", "0229", "0331", "0430", "0531", "0630", "0731", "0831", "0930", "1031", "1130", "1231"), Hmax))
Ind2[1:(length(Ind2) - 1)][diff(Ind2) == 1] <- NA
Ind2 <- Ind2[!is.na(Ind2)] ### to keep only feb 29 if both feb 28 and feb 29 exists
if (Ind2[1] < Ind1[1]) {
Ind2 <- Ind2[2:length(Ind2)]
}
if (tail(Ind1, 1) > tail(Ind2, 1)) {
Ind1 <- Ind1[1:(length(Ind1) - 1)]
}
### Aggr <- NULL; iii <- 0; for(kkk in 1:length(Ind1)) {
### iii <- iii+1; Aggr <- c(Aggr,rep(iii,length(Ind1[kkk]:Ind2[kkk]))); }
Aggr <- as.numeric(format(TmpDatesR[min(Ind1):max(Ind2)],"%Y%m")); ### more efficient
NewDatesR <- data.frame(seq(from=TmpDatesR[min(Ind1)],to=TmpDatesR[max(Ind2)],by="months"))
}
if (NewTimeFormat == "yearly") {
Ind1 <- which(format(TmpDatesR, "%m%d%H") %in% paste0(YearFirstMonthTxt, "0100"))
Ind2 <- which(format(TmpDatesR, "%m%d%H") %in% paste0(YearLastMonthTxt, NDaysInMonth[[YearLastMonth]], Hmax))
Ind2[1:(length(Ind2) - 1)][diff(Ind2) == 1] <- NA
Ind2 <- Ind2[!is.na(Ind2)]
### to keep only feb 29 if both feb 28 and feb 29 exists
if (Ind2[1] < Ind1[1]) {
Ind2 <- Ind2[2:length(Ind2)]
}
if (tail(Ind1, 1) > tail(Ind2, 1)) {
Ind1 <- Ind1[1:(length(Ind1) - 1)]
}
Aggr <- NULL
iii <- 0
for (kkk in 1:length(Ind1)) {
iii <- iii + 1
Aggr <- c(Aggr, rep(iii, length(Ind1[kkk]:Ind2[kkk])))
}
### Aggr <- as.numeric(format(TmpDatesR[min(Ind1):max(Ind2)],"%Y")); ### not working if YearFirstMonth != 01
NewDatesR <- data.frame(seq(from = TmpDatesR[min(Ind1)], to = TmpDatesR[max(Ind2)], by = "years"))
}
##_Aggreation_and_export
NewTabSeries <- data.frame(NewDatesR)
for (iCol in 2:ncol(TabSeries)) {
AggregData <- aggregate(TabSeries[min(Ind1):max(Ind2), iCol],
by = list(Aggr),
FUN = ConvertFun[iCol - 1], na.rm = FALSE)[, 2]
NewTabSeries <- data.frame(NewTabSeries, AggregData)
}
names(NewTabSeries) <- names(TabSeries)
return(NewTabSeries)
}