CODEWITHSUNDEEP
rnafillnetcdf
Dipu
Dipu

Reputation: 123

Fill value in NETCDF in R

I want R to know that this stack NetCDF contains a value -1e30 which is a fill value and it's a NA. Since the other actual values are quite similar to the NA value (1e17 to -4e18).R should understand that -1e30 is an NA and skips NA in further analysis for code na.rm=TRUE. I am working on list of files in NetCDF DATA https://drive.google.com/open?id=14OX9JAe7oZ7xPRuIzR7YB2P3OHC3zjJU

> ras <- list.files("filepath", pattern = "\\.nc$", full=TRUE)
> Data <- stack(ras)
> u2 <- mean(Data, na.rm=TRUE).

How do I write the code?

> NA<- -1e30
> Mean <-mean(Data,na.rm=TRUE)

or

> Data[Data < -1e30] = NA
> Mean <-mean(Data,na.rm=TRUE)

Will this code make R understand that -1e30 is an NA in the dataset and skip the -1e30 in taking out mean?

> dput(Data[1:100, 1:100])
NaN, 4748499736330240, NA, NA, NA, -2119029758099456, -1656679481475072, 3074375171440640, 2699225347391488, 1389911546527744, 3767667181748224,..

Upvotes: 0

Views: 543

Answers (2)

ThomasIsCoding
ThomasIsCoding

Reputation: 101383

Maybe you should used | for condition in ifelse()

r <- mean(ifelse(Data < -1e30 | is.nan(Data), NA, Data),na.rm = T)

Example with dummy data

Data <- c(NaN, 4748499736330240, NA, NA, NA, -2119029758099456, -1656679481475072, 3074375171440640, 2699225347391488, 1389911546527744, 3767667181748224, -1e31)
r <- mean(ifelse(Data < -1e30 | is.nan(Data), NA, Data),na.rm = T)

such that

> r
[1] 1.700567e+15

where

> ifelse(Data < -1e30 | is.nan(Data), NA, Data)
 [1]            NA  4.748500e+15            NA            NA
 [5]            NA -2.119030e+15 -1.656679e+15  3.074375e+15
 [9]  2.699225e+15  1.389912e+15  3.767667e+15            NA

Upvotes: 0

Georgery
Georgery

Reputation: 8117

You have a vector

a <- c(1,2,3,1e30)

You need to replace the values in there:

a <- ifelse(a == 1e30, NA, a)

Upvotes: 2

Related Questions