[R-sig-Geo] selection of data subsets in spatial classes

[Edzer J. Pebesma]

Thanks for this. For the record:

sp has a subset function for SpatialPixelsDataFrame,
but it is not exported (nor documented, I believe).
In addition, it seems not to work:

SP27GTIF_c <- sp:::subset.SpatialPixelsDataFrame(SP27GTIF_a, band1 < 100)
Error in sp:::subset.SpatialPixelsDataFrame(SP27GTIF_a, band1 < 100) :
        object "band1" not found
--
Edzer


Roger Bivand wrote:
> On Wed, 7 Jun 2006 karl.sommer at dpi.vic.gov.au wrote:
>
>   
>> Hello list
>>
>> I have been trying to figure out if it is possible to select a subset of
>> data in a SpatialGridDataFrame without prior conversion to a data.frame
>> using the method as.data.frame().
>>
>> I have imported an ESRI " *. bil " file containing different spectral bands
>> using rgdal.  The file imported as a SpatialGridDataFrame.  The data
>> originates from an aerial photograph of a row crop and I would like to
>> select pixels from within a row as opposed to those from the inter-row
>> space for further interpolation.   Inter-row pixels have a different
>> signature and therefore may be potentially screened out.  However, I have
>> been unable to access the SpatialGridDataFrame directly with, for example,
>> a  subset(x, band5 < lowerLimit, select c(a, b) ) call for selecting values
>> according to given criteria.
>>     
>
> This is what the SpatialPixelsDataFrame class is for (unless you want to 
> change resolution too):
>
> library(rgdal)
> SP27GTIF <- readGDAL(system.file("pictures/SP27GTIF.TIF", package = 
>   "rgdal")[1])
> summary(SP27GTIF)
> image(SP27GTIF, col=grey(1:9/10))
> SP27GTIF_a <- as(SP27GTIF, "SpatialPixelsDataFrame")
> summary(SP27GTIF_a)
> SP27GTIF_b <- SP27GTIF_a[SP27GTIF_a$band1 < 100,]
> summary(SP27GTIF_b)
> image(SP27GTIF_b)
>
> because SpatialPixelsDataFrame objects have their coordinates recorded 
> explicitly, and "know where they are" on a full grid. 
>
> fullgrid(SP27GTIF_b) <- TRUE
> summary(SP27GTIF_b)
>
> will drop them again, inserting NAs where band1 >= 100.
>
> This class is borrowed from the Terralib "CELL" data structure, part 
> raster, part DB record, only recording cells with data.
>
> Roger
>
>   
>> Am I using the wrong strategy, should I just demote the spatial class using
>> method as.data.frame(), do the manipulation and then promote the screened
>> values back to a spatial class?
>>
>> Regards
>>
>> Karl
>> _________________________________
>> Karl J Sommer,
>> Department of Primary Industries,
>> PO Box 905
>> Mildura, VIC, Australia 3502
>>
>> Tel: +61 (0)3 5051 4390
>> Fax +61 (0)3 5051 4534
>>
>> Email:     karl.sommer at dpi.vic.gov.au
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at stat.math.ethz.ch
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>
>>     
>
>