[R] problem with loop to put data into array with missing data forsome files

Mon Nov 20 14:49:23 CET 2006

Hi Jenny

I guess this is a silly question so sorry in advance. But why don't you
just re-download the file that failed to download completely?

John

---

John Seers
Institute of Food Research
Norwich Research Park
Colney
Norwich
NR4 7UA

tel +44 (0)1603 251497
fax +44 (0)1603 507723
e-mail john.seers at bbsrc.ac.uk                         
e-disclaimer at http://www.ifr.ac.uk/edisclaimer/ 

Web sites:

www.ifr.ac.uk   
www.foodandhealthnetwork.com

-----Original Message-----
From: r-help-bounces at stat.math.ethz.ch
[mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of Jenny Barnes
Sent: 20 November 2006 13:03
To: r-help at stat.math.ethz.ch
Subject: [R] problem with loop to put data into array with missing data
forsome files

Dear R-help community,

My main goal of this message is to find a way of skipping a file of a
month/year 
in a loop that does not exist (and making it's output into an data.out
array 
would be NA) and moving onto the next year/month in the loop to carry on
filling 
data.out with real precipitation data. 

The situation so far:
I downloaded 50 years worth of GRIB data files from the NCEP data site
http://nomad3.ncep.noaa.gov/pub/reanalysis-1/month/grb2d.gau/

I then created a loop in R to read each month of each of the 50 years
worth of 
files and only extract the precipitation records using wgrib and grep as
show in 
the code at the end of this message. I had to use grep to extract the
PRATE 
records each time, as unfortunately with these grib files precipitation
is not 
always the same record number each time. 

This worked fine until Feb 1972 (which is file 
/home/jmb/sst_precip/gribdata/197202.grib). This file did not download 
completely and as a result the file that saved to my computer did not
include 
the record for precipitation - only the first 13 records. At this point
in the 
loop it just stopped with error message:
Error in seq.default(lat.beg, lat.end, length = ny) : 
        length must be non-negative number
Which is obviously because ny does not actually exist for this file as
it cannot 
be extracted from source as this record does not exist for this file!

My question is:
Does anybody have any suggestions as to how I could code it so that my
loop will 
allow this file to just output NA into the data.out array in
data.out$data and 
maybe keep FALSE in data.out$date for the month&years where this has
occured and 
then continue to loop around all the other months and years in
succession? 

Things that don't work that I have tried:
1)I cannot just restart the loop after any anomalous years as this
"resets" 
data.out and changes its dimensions depending on the number of years
looped 
(nyrs) etc.
2)The first point of trouble is when I call upon ny in the line
>lats <- seq(lat.beg, lat.end, length=ny)
as ny is NA in this record - so I tried putting the line
>test <- dims[15]
>if(test != "NA"){
with the closing } at the end of the rest of the loop
...this gives the error message
>Error in if (test != "NA") { : missing value where TRUE/FALSE needed
3) My thought was that maybe I would need something like
>if (parameter == "PRATE" ){
then continue with loop; but this will obviously not work by itself and
gives 
the error message
>Error in if (test != "NA") { : missing value where TRUE/FALSE needed
- would need to specify what to do if parameter is not =="PRATE". I
encourage 
anybody with another idea to contact me with their suggestions!

Thank you for taking the time to read my problem,

Sincerely,

Jennifer Barnes

r-help at stat.math.ethz.ch
Here is my code:

#takes precip data for every year and month needed
#from file saved on computer already
#and sorts it into a list (array)

#set some control parameters

startyr <-1954 #note needs to be year before data wanted
nyrs <- 50
months <- nyrs*12
k <- 0
library(chron)
library(utils)
data.out <- list(lats=seq(88.542, -88.542, length=94),
		  lons=seq(0, 360-1.875, length=192),
		  date=vector(length=months),
		  data=array(NA, c(months, 94*192))
) 
#the loop
for(yr in 1:nyrs){
     for(mon in 1:12){
     	i<- startyr+yr

	if(mon < 10) {j <- 0}
	if(mon < 10) {mon2 <- mon}
	if(mon == 10) {(j <- 1)}
	if(mon == 10) {mon2 <- 0}
 	if(mon == 11) {(j <- 1)}
	if(mon == 11) {mon2 <- 1}
	if(mon == 12) {(j <- 1)}
	if(mon == 12) {mon2 <- 2}

	outfile <- paste("/home/jmb/sst_precip/gribdata/",i, j, mon2,
".grib", 
sep="")  
		#decode the grib 
	cmd <- paste("/usr/local/bin/wgrib ", outfile, "| grep
\":PRATE:\" | 
/usr/local/bin/wgrib", outfile, "-i -V -text -o /tmp/wgrib.output")
	status <- system(cmd, intern=T)

	#extract metadata needed for working with julian dates:
	dt <- unlist(strsplit(status[1], " "))[3]
	dt.yr <- as.numeric(substr(dt, 1,4))
	dt.mon <- as.numeric(substr(dt, 5,6))
	dt.day <- as.numeric(substr(dt, 7,8))
	dt.hr <- as.numeric(substr(dt, 9,10))
	dt.jul <- julian(dt.mon, dt.day+(dt.hr/24), dt.yr)

	parameter <- unlist(strsplit(status[1], " "))[4]

	param <- status[2]
	dims <- unlist(strsplit(status[3], " "))
	nx <- as.numeric(dims[13])
	ny <- as.numeric(dims[15])
	temp <- unlist(strsplit(status[5]," "))
	lat.beg <- as.numeric(temp[6])
	lat.end <- as.numeric(temp[8])
	lats <- seq(lat.beg, lat.end, length=ny)
	temp <- unlist(strsplit(status[6]," "))
	lon.beg <- as.numeric(temp[14])
	lon.end <- as.numeric(temp[16])
	if(lon.beg > lon.end){lon.beg <- (lon.beg -360)} #since sp
prefers -180 
to 180 encoding 
	lons <- seq(lon.beg, lon.end, length=nx)
	lons[lons > 180] <- lons[lons > 180]-360	 #since sp
prefers -180 
to 180 encoding 

	#read flat file into R (header first)
	dims <- scan("/tmp/wgrib.output", nlines=1, quiet=T)
	nrec <- dims[1]*dims[2]
	indata <- paste("indata.", i, j, mon2, sep="")
	indata <- scan("/tmp/wgrib.output", 'numeric', nlines=nrec,
skip=1, 
quiet=T)
	indata <- matrix(as.numeric(indata), nrow=dims[1], ncol=dims[2],

byrow=F)

	#to get precip into mm/month, not mm/sec:
	if((dt.mon==02)&(dt.yrendnumber%%4==0)){days.mon <- 29} else
{days.mon 
<-28}    #to take into account leap-years!!!
	if(dt.mon==01){days.mon <- 31}
	if(dt.mon==03){days.mon <- 31}
	if(dt.mon==05){days.mon <- 31}
	if(dt.mon==07){days.mon <- 31}
	if(dt.mon==08){days.mon <- 31}
	if(dt.mon==10){days.mon <- 31}
	if(dt.mon==12){days.mon <- 31}
	if(dt.mon==04){days.mon <- 30}
	if(dt.mon==06){days.mon <- 30}
	if(dt.mon==09){days.mon <- 30}
	if(dt.mon==11){days.mon <- 30}
	sec.mon <- days.mon*24*60*60

	newindata <- paste("newindata.",i, j, mon2, sep="")
	newindata <- sec.mon*indata

	k<-k+1
	data.out$date[k] <- paste(i, j, mon2, sep="")
	data.out$data[k,] <- newindata
}
}

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Jennifer Barnes
PhD student - long range drought prediction
Climate Extremes
Department of Space and Climate Physics
University College London
Holmbury St Mary, Dorking
Surrey
RH5 6NT
Web: http://climate.mssl.ucl.ac.uk

______________________________________________
R-help at stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.