Skip to content

Request for help about running a loop and reading .NC files

4 messages · Bhaskar Mitra, Ivan Krylov, Bert Gunter +1 more

Original
Bhaskar Mitra
# 
Hello Everyone,

I have written a loop which reads hundreds of .nc files
and extract information from each .nc file and
exports that corresponding information as a csv file.

The loop works fine until it encounters a .nc file which it cannot read
and the loop stops.

 I would appreciate if anyone can suggest how can I modify
the loop, whereby the loop will continue to run by bypassing those
particular
files which it cannot read or if any particular file has an error.

In the end, I was also hoping to modify the loop such that it
will generate a report which will inform me which
files were not read by the loop.  The codes are given below

Thanks for your help,

Regards,
Bhaskar Mitra


#_------------------------------------------------------------------


library(ncdf4)
library(reshape2)
library(dplyr)
library(stringr)


setwd("directory path")

Output <- "directory path"

flist <- list.files(path ="NCFiles/", pattern = "^.*\\.(nc|NC|Nc|Nc)$")

for (i in 1: length(flist))
{

  nc <- nc_open(paste0("NCFiles/",flist[i]))
  mean1 <- ncvar_get(nc,attributes(nc$dim)$names[3])
  nc_close(nc)

  mean_chl_df <- melt(mean1)
  trial.table.df <-as.data.frame(mean_chl_df)

write.csv(trial.table.df,paste0(Output,"/",tools::file_path_sans_ext(flist[i]),".csv"))

}
Ivan Krylov
# 
On Mon, 23 Mar 2020 16:16:42 -0700
Bhaskar Mitra <bhaskar.kolkata at gmail.com> wrote:

            

      
      
      
    

        
See ?try and ?tryCatch. With `try`, you can use inherits(x,
'try-error') to check whether execution failed. With `tryCatch` you can
choose your own value to return instead of the error object.

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Bert Gunter
    

    

      
      #
    

  


  

      
?tryCatch

See also:
https://www.r-bloggers.com/error-handling-in-r/


Bert Gunter

"The trouble with having an open mind is that people keep coming along
and sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )

        
On Tue, Mar 24, 2020 at 5:13 AM Bhaskar Mitra <bhaskar.kolkata at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      David W. Pierce
    

    

      
      #
    

  


  

      
Hey Bhaskar,

Dave Pierce here, the author and maintainer of the ncdf4 package. What
platform are you working on? If you work on Linux, I'd be happy to send you
a new source tar version of the ncdf4 package I've made that handles this
case. It now has a settable parameter in nc_open called "return_on_error",
and if you set that to TRUE, then nc_open always returns. If no error was
encountered, then the return list element $error is FALSE, and if an error
was encountered, $error is set to TRUE. So you could do this:

for( ifile in 1:nfiles ) {
    nc = nc_open( filename[ifile], return_on_error=TRUE )
    if( nc$error == FALSE ) {
        ... do your processing here ...
        }
  else
        print(paste("Error encountered, skipping file", filename[ifile]))

Regards,

--Dave

On Tue, Mar 24, 2020 at 5:13 AM Bhaskar Mitra <bhaskar.kolkata at gmail.com>
wrote: