[Rd] Bug found in cut.Date, solution found

Duncan Murdoch murdoch@dunc@n @end|ng |rom gm@||@com
Fri Feb 5 23:47:53 CET 2021


If you don't hear something from someone in R Core, you should submit 
this as a bug report on bugs.r-project.org.  (You might not have an 
account there; if you don't, someone from R Core will have to give you 
one.  Feel free to remind them if nothing happens in a few days!)

Duncan Murdoch

On 05/02/2021 4:01 p.m., Christopher Carbone wrote:
> Greetings,
> 
> I found a bug in the cut.Date function of base R. My session info is
> at the bottom.
> 
> The bug is in the "months" section of the code. Consider the following:
> 
> # Create vector with 365 days of year and cut into 12 months
> dates <- seq(as.Date("2021-1-1"),as.Date("2021-12-31"),by="day")
> months <- cut(dates,"months",labels=1:12)
> table(months)
> # months
> # 1  2  3  4  5  6  7  8  9 10 11 12
> # 31 28 31 30 31 30 31 31 30 31 30 31
> # cut.Date works perfectly
> 
> # Extend dates vector by adding following January and cut into 13 months
> dates <- seq(as.Date("2021-1-1"),as.Date("2022-1-31"),by="day")
> months <- cut(dates,"months",labels=1:13)
> # Error in cut.default(unclass(x), unclass(breaks), labels = labels,
> right = right,  :
> #                        lengths of 'breaks' and 'labels' differ
> months <- cut(dates,"months",labels=1:14)
> table(months)
> # months
> # 1  2  3  4  5  6  7  8  9 10 11 12 13 14
> # 31 28 31 30 31 30 31 31 30 31 30 31 31  0
> # Bug: cut.Date adds a superfluous bin
> # This is because February has less than 31 days
> # Thus, the code adjusts the end into March instead of February
> 
> Inspecting the source code, I see the problem resides in lines 31
> through 41, the section dealing with breaks="months":
> 
>      if (valid == 3L) {
>        start$mday <- 1L
>        start$isdst <- -1L
>        end <- as.POSIXlt(max(x, na.rm = TRUE))
>        step <- if (length(by2) == 2L)
>          as.integer(by2[1L])
>        else 1L
> 38  end <- as.POSIXlt(end + (31 * step * 86400))
>        end$mday <- 1L
>        end$isdst <- -1L
>        breaks <- as.Date(seq(start, end, breaks))
> 
> When the end is adjusted in line 38, there are instances where an
> unnecessary extra bin is created.
> This is accounted for in the breaks="quarters" case by testing at the
> end and removing the final break if necessary.
> The same method employed here fixes the bug:
> 
>      if (valid == 3L) {
>        start$mday <- 1L
>        start$isdst <- -1L
>       maxx <- max(x, na.rm = TRUE)   # Added line
>       end <- as.POSIXlt(maxx)             # Modified line
>        step <- if (length(by2) == 2L)
>          as.integer(by2[1L])
>        else 1L
>        end <- as.POSIXlt(end + (31 * step * 86400))
>        end$mday <- 1L
>        end$isdst <- -1L
>        breaks <- as.Date(seq(start, end, breaks))
>       lb <- length(breaks)                   # Added line
>       if (maxx < breaks[lb - 1])           # Added line
>         breaks <- breaks[-lb]               # Added line
> 
> I modified cut.Date with the above code changes and stored it in a
> function called cut_Date():
> 
> dates <- seq(as.Date("2021-1-1"),as.Date("2022-1-31"),by="day")
> months <- cut_Date(dates,"months",labels=1:13)
> table(months)
> # months
> # 1  2  3  4  5  6  7  8  9 10 11 12 13
> # 31 28 31 30 31 30 31 31 30 31 30 31 31
> 
> Thanks for your time and consideration!
> 
>> sessionInfo()
> R version 4.0.3 (2020-10-10)
> Platform: x86_64-w64-mingw32/x64 (64-bit)
> Running under: Windows 10 x64 (build 19041)
> 
> Matrix products: default
> 
> locale:
> [1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252
> [3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C
> [5] LC_TIME=English_United States.1252
> 
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
> 
> loaded via a namespace (and not attached):
> [1] compiler_4.0.3 tools_4.0.3    tinytex_0.28   xfun_0.19
> 
> Many Blessings,
> Christopher Carbone
> 
> "Follow your bliss and doors will open where you didn't know they were
> going to be..."
> -Joseph Campbell
> 
> ______________________________________________
> R-devel using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>



More information about the R-devel mailing list