[R-sig-Geo] rasterVis levelplot: Use colortable in raster with attribute table

[Gregor Ratzmann]

Thanks for your reply, Oscar,

I'm sorry if the code appears somewhat confusing so I try to explain what it is supposed to do.

So if I use the example from the ratify function help:

r <- raster(nrow=10, ncol=10)
r[] = 1
r[51:100] = 2
r[3:6, 1:5] = 3
r <- ratify(r)

rat <- levels(r)[[1]]
rat$landcover <- c('Pine', 'Oak', 'Meadow')
rat$code <- c(12,25,30)
levels(r) <- rat

cols=c("red","green","blue")

levelplot(r,col.regions=cols)


everything works fine as expected.

So I will try to explain what my code is supposed to do. I have a thematic raster with attribute table. The raster takes on values between 1 and 28 each value representing one class. The value and corresponding class name are stored in the attribute tables column 5 and 4, respectively.

#First I load the raster with the attribute table and color table

r<-raster(##some raster with attribute and colortable##))

#then I create a data frame from the levels of the thematic raster

ratdf<-as.data.frame(levels(r))


#now I extract column 5 of the RAT df which holds the IDs

rat<-as.data.frame(ratdf[[5]])

#Next step: create a vector with the classnames from RAT df col 4

nvec<-as.character(ratdf[[4]])

#then join the class names vector with the ID data frame 

rat$classes<-nvec

#assign the column names of rat 
names(rat)<-c("ID","classes")

#create an empty raster r1 from the r template

r1<-raster(r)

#assign the values of r to r1

r1[]<-r[]

#make a thematic raster of r1
 
r1<-ratify(r1)

#assign the levels to r1 from rat

levels(r1)<-rat


#next steps are to create a color ramp with only the valid colors (since the original raster is 8 bit it could contain up to 256 #colors but I only need colors 2 to 29) and plot the thing

cols<-r at legend@colortable
cols<-cols[2:29]

levelplot(r1,col.regions=cols)

####


What puzzles me is that after I created the new thematic raster with ratify, and I call the raster, it is displayed but under attributes appears the error: Error in 1:ncol(r) : argument of length 0. Maybe this has something to do with the problems plotting that raster. Btw, the same error appears as well when I use as.factor instead of ratify...

Cheers

greg

________________________________________
Von: oscar.perpinan at gmail.com [oscar.perpinan at gmail.com]" im Auftrag von "Oscar Perpiñan [oscar.perpinan at upm.es]
Gesendet: Freitag, 7. Februar 2014 19:40
An: Gregor Ratzmann
Cc: r-sig-geo at r-project.org
Betreff: Re: [R-sig-Geo] rasterVis levelplot: Use colortable in raster with attribute table

Hello,

> I am trying to plot a raster which comes with an attribute table and a colortable using the rasterVis levelplot function.
>
> Here is my first question: is it possible to address the colortable slot directly in the levelplot function (as is possible for the plot3D through the useLegend argument)?

Not yet. It is still in my To-do list.

> Since I did not find a way to do it, I tried the below. Note that the original attribute table is not arranged as is neccessary for levelplot to read it so I first create a new raster and assign to it the attributes which are needed:
> I know the procedure is improvable but for now I think it should do. The problem: Everything is plotted fine as expected, except that the last two classes of the raster are assigned the same color. Since there are the same number of classes as number of colors I do not really understand why this happens and, strange enough, in the color key all colors are plotted fine with the correct labels....additionally, when i do it using the plot() function it is working fine. When I used a different color ramp for levelplot it worked out for all classes and everything was well differentiated... Is there maybe some hex color code which levelplot can't read? Please appologize if I didn't express clearly what I want but I am still at the beginning of using R.

I don't understand what you are trying to do with your code. Could you
provide a reproducible example? Try the examples of the ratify help
page.

Best,

Oscar.