Qfyilyi

I am using the following code to check P-values of a linear trend, but it seems that the loop is not working properly as I cannot see a 2-D map of P-value but only a row

library(chron)

library(RColorBrewer)

library(lattice)

library(ncdf4)

#-------------------------------------------------------------------------------------------

options(warn=-1)



ncin <- nc_open("MOD04_10K_Winter.nc", readunlim=FALSE)

#print(ncin)

lon <- ncvar_get(ncin, varid="Longitude", start=NA, count=NA, verbose=FALSE,

          signedbyte=TRUE, collapse_degen=TRUE, raw_datavals=FALSE )

lat <- ncvar_get(ncin, varid="Latitude", start=NA, count=NA, verbose=FALSE,

                signedbyte=TRUE, collapse_degen=TRUE, raw_datavals=FALSE )

aod <- ncvar_get(ncin, varid="AOD", start=NA, count=NA, verbose=FALSE,

                signedbyte=TRUE, collapse_degen=TRUE, raw_datavals=FALSE )



px           <- matrix(nrow = 1:length(lon), ncol = 1:length(lat))

is.matrix(px)



for (lo in 1:length(lon)) {

  for (la in 1:length(lat)) {

   int1a    = aod[lo, la,]



   # if mean of int is finite then proceed else fill NA to all arrays

   mn = mean(int1a, trim = 0, na.rm = FALSE)

   if (is.finite(mn)) 

    {

    print("---------------- Reading Finite data -------------")

    xs       = 1:30                                     

    fn1a     = lm(int1a~xs)       # Function_NCP

    p_val    = summary(fn1a)$coefficients[2, 4]          # Saving p-value

    if (p_val < 0.05) {print("statisticlly significant")} else {print("statisticlly in-significant")}

    print(p_val)

    print(lo)

    print(la)

    px[lo][la] = p_val                                   # variables in  only (?)

   } 



  }                                                      # latitude dimension

}   

If I am using [lo, la] instead of [lo][la] I am having the following error

Error in [<-(*tmp*, lo, la, value = 0.0543481042240582) :

subscript out of bounds

Sorry if the solution is very trivial, I have just started working in R.

You have just to make a small fix on the matrix px declaration. Now you set the number of rows and columns as vectors: nrow = 1:length(lon) and nrow = 1:length(lon). R silently takes only the first elements of these vectors and generates a 1 to 1 matrix. (Actually, it would generate a warning, by the warnings are supressed!)

So, the solution is

px <- matrix(nrow = length(lon), ncol = length(lat))