Skip to content

Error in as.POSIXlt.character(x, tz, ...)

7 messages · Ivan Krylov, Jeff Newmiller, Jinsong Zhao

Original
Jinsong Zhao
# 
Hi there,

When I run the following code in R 4.3.0 on FreeBSD, I got error.

 > as.POSIXct("1970-01-01 00:00.00 UTC")
Error in as.POSIXlt.character(x, tz, ...) :
   character string is not in a standard unambiguous format

The same code could give correct answer in R 4.3.0 on Windows, and R 
4.1.2 on a old CentOS.

 > as.POSIXct("1970-01-01 00:00.00 UTC")
[1] "1970-01-01 CST"

I confirmed that all the 3 OSes give the same time zone:
 > Sys.timezone()
[1] "Asia/Shanghai"

What's wrong with the R on FreeBSD? Any hint? Thanks in advance.

Best,
Jinsong
Ivan Krylov
# 
? Thu, 11 May 2023 16:56:41 +0800
Jinsong Zhao <jszhao at yeah.net> ?????:
Do strptime("1970-01-01 00:00.00 UTC", "%Y-%m-%d %H:%M", '') and
strptime("1970-01-01 00:00.00 UTC", "%Y-%m-%d", '') both return NA for
you on FreeBSD? If not, can you trace the execution using R's debugger
to see which parts are returned as NA, failing the test in
as.POSIXlt.character?

I think that strptime() should have discarded the trailing parts of the
string and matched at least the beginning (though not the ".00 UTC"
part because it doesn't correspond to the formats tried by strptime()).
Moreover, there shouldn't be significant difference between the
platforms, because R seems to always use an internal implementation of
strptime(). Are you using a multi-byte locale? (What's your
sessionInfo() on both a system where it works and the system where it
fails?)
Jeff Newmiller
# 
I find your claim suspect... a period is not interchangeable with a colon.
On May 11, 2023 1:56:41 AM PDT, Jinsong Zhao <jszhao at yeah.net> wrote:

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Jinsong Zhao
    

    

      
      #
    

  


  

      
On 2023/5/11 17:24, Jeff Newmiller wrote:

            

      
      
      
    

        
Good catching, but

In R 4.3.0 on FreeBSD, the same error after changing the period to a colon:
 > as.POSIXct("1970-01-01 00:00:00 UTC")
Error in as.POSIXlt.character(x, tz, ...) :
   character string is not in a standard unambiguous format

I copy the code from a web page. It works in R 4.3.0 on Windows and R 
4.1.2 on an old CentOS.


Best,
Jinsong

            

      
      
      
    

        

      
      
    

            

      
      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Jinsong Zhao
    

    

      
      #
    

  


  

      
On 2023/5/11 17:22, Ivan Krylov wrote:

            

      
      
      
    

        
Both codes do not return NA on FreeBSD:

 > strptime("1970-01-01 00:00.00 UTC", "%Y-%m-%d %H:%M", '')
[1] "1970-01-01 CST"
 > strptime("1970-01-01 00:00.00 UTC", "%Y-%m-%d", '')
[1] "1970-01-01 CST"

I do not know how to use R's debugger...

            

      
      
      
    

        
The sessionInfo on FreeBSD said:
tzcode source: system (FreeBSD)

But on Windows it said:
tzcode source: internal

Could it be because of this?


The sessionInfo on FreeBSD is:

 > sessionInfo()
R version 4.3.0 (2023-04-21)
Platform: amd64-portbld-freebsd13.1 (64-bit)
Running under: FreeBSD xw6600 13.2-RELEASE FreeBSD 13.2-RELEASE 
releng/13.2-n254617-525ecfdad597 GENERIC amd64

Matrix products: default
LAPACK: /usr/local/lib/R/lib/libRlapack.so.4.3.0;  LAPACK version 3.11.0

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

time zone: Asia/Shanghai
tzcode source: system (FreeBSD)

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

loaded via a namespace (and not attached):
[1] compiler_4.3.0

The sessionInfo on Windows is:
 > sessionInfo()
R version 4.3.0 (2023-04-21 ucrt)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 11 x64 (build 22621)

Matrix products: default


locale:
[1] LC_COLLATE=Chinese (Simplified)_China.utf8
[2] LC_CTYPE=Chinese (Simplified)_China.utf8
[3] LC_MONETARY=Chinese (Simplified)_China.utf8
[4] LC_NUMERIC=C
[5] LC_TIME=Chinese (Simplified)_China.utf8

time zone: Asia/Shanghai
tzcode source: internal

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

loaded via a namespace (and not attached):
[1] compiler_4.3.0 tools_4.3.0
 >

The sessionInfo on CentOS is:
 > sessionInfo()
R version 4.1.2 (2021-11-01)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)

Matrix products: default
BLAS:   /home/jszhao/software/lib64/R/lib/libRblas.so
LAPACK: /home/jszhao/software/lib64/R/lib/libRlapack.so

locale:
  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C
  [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8
  [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8
  [7] LC_PAPER=en_US.UTF-8       LC_NAME=C
  [9] LC_ADDRESS=C               LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

loaded via a namespace (and not attached):
[1] compiler_4.1.2
 >

  
  


  


      

    

    

      
      

        


  

        

      Jinsong Zhao
    

    

      
      #
    

  


  

      
The origin of this problem is when I try to update rstan package in R 
4.3.0 on FreeBSD. I get the same error as that 
https://github.com/stan-dev/rstan/issues/612

However, I did not encounter this problem when using R 4.2.x or 4.1.x on 
FreeBSD.

Best,
Jinsong

        
On 2023/5/11 16:56, Jinsong Zhao wrote:

            

      
      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Ivan Krylov
    

    

      
      #
    

  


  

      
? Thu, 11 May 2023 17:51:33 +0800
Jinsong Zhao <jszhao at yeah.net> ?????:

            

      
      
      
    

        
I must have misread the code of as.POSIXlt.character then.

            

      
      
      
    

        
In short, debugonce(as.POSIXlt.character) will prepare R to start the
debugger once the function is called. The debugger prompt is described
in ?browser; it's mostly like R's prompt (you can call print(variable)
and other, more complicated expressions), but there are special
commands like "n" (next statement), "s" (step inside the function call)
and so on.

Call:

debugonce(as.POSIXlt.character)
as.POSIXct("1970-01-01 00:00.00 UTC")

Use n to step over function calls and s to step inside for loops. See
what's returned by strptime() calls inside this function. It's raising
an exception because all of them contain NAs somewhere; which element is
NA? What are the exact arguments used to call strptime()?

This doesn't seem to be due to differences in locale: both CentOS and
FreeBSD use en_US.UTF-8.