[R-sig-Geo] Error with disaggregate

Ned Horning horning at amnh.org
Tue Feb 18 19:42:27 CET 2014


Robert,

I downloaded the development version and now it's working fine - thanks.

Ned

On 02/18/2014 01:12 PM, Robert J. Hijmans wrote:
> Ned,
>
> I see nothing wrong with your code. Works for me on Win7.  Perhaps
> because of a fix in the chunk size (blockSize) computation in the
> development version.
> Can you check if it works with that version?
>
>   install.packages("raster", repos="http://R-Forge.R-project.org")
>
>
> I do not understand the error message though, as you should be writing
> fewer bytes then 2^31
>> 2^31 / (8*ncell(c))
> [1] 4.532697
>
> This may go away if you add a filename, and perhaps a datatype (FLT4S)  argument
>
> c <- disaggregate(b, fact=32, filename='dis.tif', datatype='FLT4S')
>
> Or if you set
> rasterOptions(chunksize= ) temporarily to something lower.
>
> I note that I should add a "block_fun" method to do this in one step
>
> Robert
>
> On Tue, Feb 18, 2014 at 9:47 AM, Ned Horning <horning at amnh.org> wrote:
>> Hi -
>>
>> I am trying calculate the sum of blocks of Landsat pixels (30m resolution)
>> that fit within MODIS pixels 960m resolution) and store that values at the
>> Landsat cell size. To accomplish that I am running aggregate with the "sum"
>> function on a Landsat band and then I run disaggregate on the result to get
>> the resolution back to the Landsat pixel size. When I do this the
>> disaggregate function gives me the following error:
>> --
>> Error in writeBin(v, x at file@con, size = x at file@dsize) :   only 2^31-1 bytes
>> can be written in a single writeBin() call
>> --
>>
>> I can recreate the error with this script:
>> --
>> a <- raster(nrows=7776, ncols=7616, xmn=718335, xmx=946815, ymn=-757485,
>> ymx=-524205, crs="+proj=utm +zone=35 +datum=WGS84 +units=m +no_defs
>> +ellps=WGS84 +towgs84=0,0,0")
>> a[] <- 1
>> b <- aggregate(a, fact=32, fun=sum)
>> c <- disaggregate(b, fact=32)
>> --
>>
>> I am running R 3.0.2 on Ubuntu 12.04 with version 2.2-12 of the raster
>> package.
>>
>> Please let me know if you see any problems with my code.
>>
>> Thanks.
>>
>> Ned
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo



More information about the R-sig-Geo mailing list