0

I have many NCDF files in a folder. I try to extract them into raster brick using raster and ncdf4 packages. If I separately extract each NCDF file it works. However, I try to extract all files using for loop then it gives me error.

R<-list.files("D:/Results/TimeSeries/NETCDF/")
r<-brick(paste0("D:/Results/TimeSeries/NETCDF/",R[[1]]),varname="T_min")

for(i in 2:length(R)){
r1<-brick(paste0("D:/Results/TimeSeries/NETCDF/",R[[i]]),varname="T_min")
r<-brick(r,r1)
}

Error in as.integer(nl) : cannot coerce type 'S4' to vector of type 'integer'

If I look at r and r1 separately they seem to have same extent and both are raster brick type:

> r
class       : RasterBrick 
dimensions  : 81, 81, 6561, 122  (nrow, ncol, ncell, nlayers)
resolution  : 1, 1  (x, y)
extent      : 0.5, 81.5, 0.5, 81.5  (xmin, xmax, ymin, ymax)
coord. ref. : NA 
data source : D:\Results\TimeSeries\NETCDF\timeseries_1km_2026.nc 
names       : X0026.05.02, X0026.05.03, X0026.05.04, X0026.05.05, X0026.05.06, X0026.05.07, X0026.05.08, X0026.05.09, X0026.05.10, X0026.05.11, X0026.05.12, X0026.05.13, X0026.05.14, X0026.05.15, X0026.05.16, ... 
Date        : 0026-05-02, 0026-08-31 (min, max)
varname     : T_min 

> r1
class       : RasterBrick 
dimensions  : 81, 81, 6561, 122  (nrow, ncol, ncell, nlayers)
resolution  : 1, 1  (x, y)
extent      : 0.5, 81.5, 0.5, 81.5  (xmin, xmax, ymin, ymax)
coord. ref. : NA 
data source : D:\Results\TimeSeries\NETCDF\timeseries_1km_2027.nc 
names       : X0027.05.02, X0027.05.03, X0027.05.04, X0027.05.05, X0027.05.06, X0027.05.07, X0027.05.08, X0027.05.09, X0027.05.10, X0027.05.11, X0027.05.12, X0027.05.13, X0027.05.14, X0027.05.15, X0027.05.16, ... 
Date        : 0027-05-02, 0027-08-31 (min, max)
varname     : T_min

Please help.

rar
  • 894
  • 1
  • 9
  • 24
  • Ok... It was a small mistake. Instead of `brick(r,r1)`, we need to use `stack(r,r1)` and it works fine. – rar Nov 03 '16 at 11:14

1 Answers1

0

There is no need to loop, raster is vectorized, try

p <- "D:/Results/TimeSeries/NETCDF"
R <- list.files(p, pattern = "nc$")

r <- raster::stack(file.path(p, R), varname = "T_min")

If you did need to loop, I'd do it like this: 

r <- raster::stack(lapply(file.path(p, R), raster::raster, varname = "T_min"))

Edit: replace raster::raster with raster::stack.

Also note the use of file.path, and the facilities available within list.files. (Pasting text for file paths can be problematic, and is more complicated than using available functions).

mdsumner
  • 29,099
  • 6
  • 83
  • 91
  • Actually I have 40 NCDF files and each has 122 timeseries data. So total I should have 4880 layers in final. However, if I use these codes I get 1st layer from each file, thus output has 40 layers. I would appreciate if this can be done without loop. – rar Nov 03 '16 at 12:21
  • Use raster::stack in the lapply, ? – mdsumner Nov 03 '16 at 12:38
  • It works with `r <- raster::stack(lapply(file.path(p, R), raster::stack, varname = "T_min"))` So we do need loop for this. But thanks for this much easier code! – rar Nov 03 '16 at 13:28
  • I forgot about the multi-variable thing, you should check the unlimited dimension values (time here), no guarantee they are regular, in order, compete, consistent units etc. I don't trust these collections generally, way too much md is spread out and complicated. – mdsumner Nov 03 '16 at 19:54