Merge pull request #29 from Nixtla/refactor/generate_output_dates

feat: vectorized date generation

R/generate_output_dates.R

@@ -14,18 +14,15 @@
 #'   dates_df <- .generate_output_dates(df_info, freq, h)
 #' }
 #'
-.generate_output_dates <- function(df_info, freq, h){
-
-  new_dates <- vector("list", nrow(df_info))
-  r_freq <- .r_frequency(freq)
-
-  for(i in 1:nrow(df_info)){
+.generate_output_dates <- function(df_info, freq, h) {
+  new_dates <- lapply(1:nrow(df_info), function(i) {
     start_date <- df_info$dates[i]
+    r_freq <- .r_frequency(freq)
 
-    if(freq %in% c("QE", "Q")){
-      dt <- seq(from = start_date, by = r_freq, length.out = h+1)
-      month <- lubridate::month(start_date)
+    if (freq %in% c("QE", "Q")) {
+      # End of quarter dates are: "YYYY-03-31", "YYYY-06-30", "YYYY-09-30", "YYYY-12-31".
       dt <- seq(from = start_date, by = "quarter", length.out = h+1)
+      month <- lubridate::month(start_date)
 
       # Calendar adjustments
       if (month %in% c(3, 12)) {
@@ -44,21 +41,25 @@
           }
         }
       }
-
-    }else if(freq %in% c("ME", "M")){
-      start_date <- start_date+lubridate::days(1)
-      dt <- seq(from = start_date, by = r_freq, length.out = h+1)-lubridate::days(1)
-    }else{
-      dt <- seq(df_info$dates[i], by = r_freq, length.out = h+1)
+      dt <- format(dt, "%Y-%m-%d")
+    } else if (freq %in% c("ME", "M")) {
+      dt <- seq(from = start_date + lubridate::days(1), by = r_freq, length.out = h+1) - lubridate::days(1)
+      dt <- format(dt, "%Y-%m-%d")
+    } else if(freq %in% c("B")){
+      dt <- seq(from = start_date, by = "day", length.out = h+1+ceiling(h/5)*2+2) # ceiling(h/5)*2+2 ~ number of weeks*2 days (Saturday and Sunday) plus an extra weekend to be on the safe side
+      dt <- dt[!weekdays(dt) %in% c("Saturday", "Sunday")]
+      dt <- format(dt, "%Y-%m-%d")
+    } else {
+      dt <- seq(from = start_date, by = r_freq, length.out = h+1)
     }
 
-    new_dates[[i]] <- dt[2:length(dt)]
-  }
+    dt[2:(h+1)]
+  })
 
   dates_df <- data.frame(lapply(new_dates, as.POSIXct))
 
   ids <- df_info$unique_id
-  if(inherits(df_info$unique_id, "numeric") | inherits(df_info$unique_id, "integer")){
+  if (inherits(df_info$unique_id, "numeric") | inherits(df_info$unique_id, "integer")) {
     ids <- as.character(df_info$unique_id)
   }
   names(dates_df) <- ids

vignettes/special-topics.Rmd

@@ -77,8 +77,9 @@ The frequency parameter is crucial when working with time series data because it
 | Hourly                    | h            | 
 | Minute-level              | min          |
 | Second-level              | s            |
+| Business day              | B            |
 
-In this table, QS and MS stand for quarter and month start, while QE and ME stand for quarter and month end. Hourly and subhourly frequencies can be preceded by an integer, such as "6h", "10min" or "30s". **Only the aliases "min" and "S" are allowed for minute and second-level frequencies**.  
+In this table, QS and MS stand for quarter and month start, while QE and ME stand for quarter and month end. Hourly and subhourly frequencies can be preceded by an integer, such as "6h", "10min" or "30s". **Only the aliases "min" and "S" are allowed for minute and second-level frequencies**. 
 
 The default value of the frequency parameter is `NULL`. When this parameter is not specified, `nixtlar` will attempt to determine the frequency of your data. 
 
@@ -89,10 +90,7 @@ fcst <- nixtlar::nixtla_client_forecast(df, h = 8, level = c(80,95)) # freq = "h
 # infer the frequency when `freq` is not specified 
 ```
 
-If you are dealing with irregular frequencies, such as business days or custom holiday calendars, you must specify them directly. For instance, for business days, you should set `freq="B"`, which corresponds to the pandas alias for business day frequency. Please refer to [pandas's offset aliases](https://pandas.pydata.org/pandas-docs/stable/user_guide/timeseries.html#timeseries-offset-aliases) for more information. 
-
-When dealing with weekly frequency (`W`), `nixtlar` assumes that the weeks start on Sunday. Consequently, it will return dates corresponding to weeks that begin on Sundays. If your weeks start on a different day, for example, Mondays, you should specify the frequency as `W-MON`. You can select any day of the week with the aliases `W-MON`, `W-TUE`, `W-WED`, `W-THU`, `W-FRI`, and `W-SAT`.
-
+**Currently, `nixtlar` can't infer business day frequency, so you must set it directly using `freq="B"`.** 
 
 ```{r, include=FALSE}
 options(original_options)