[R-sig-Geo] raster::extract NAs introduced?

[Robert J. Hijmans]

Agus,

> any(is.na(SGRGBF40))

returns a RasterLayer with

> min value   : 0
> max value   : 1

TRUE == 1, so there is at least one NA value

also see:

count(SGRGBF40, NA)

summary(SGRGBF40, maxsamp=ncell(SGRGBF40))


Best, Robert


On Fri, Mar 25, 2011 at 6:58 AM, Agustin Lobo <alobolistas at gmail.com> wrote:
> Hi
>
> I've found that extract() introduces NAs:
> SGRGBF40 = brick("/media/Iomega_HDD/UAVetal/CALIBRACIONRADIOM/TESTCASA/CALSEL/SGRGB/SGRGBWBPS125F40.tif")
> calibf2 = readOGR(dsn="/media/Iomega_HDD/UAVetal/CALIBRACIONRADIOM/TESTCASA/CALSEL",layer="calibf2",stringsAsFactors=F)
> projection(SGRGBF40) = projection(calibf2)
> plot(subset(SGRGBF40,1))
> plot(calibf2,add=T)
> (overlay is ok now)
>
>> summary(SGRGBF40)
> Cells:  4646400
> NAs  :  0 0 0
>
>            1     2     3
> Min.     2629     3     3
> 1st Qu. 34100 22200 18290
> Median  42470 27480 24680
> Mean    40900 28510 25360
> 3rd Qu. 48360 33100 29010
> Max.    65510 65510 64300
> summary based on a sample of 5000 cells, which is 0.107610192837466 %
> of all cells
> To make sure (NAs could outside the sample):
>> any(is.na(SGRGBF40))
> class       : RasterLayer
> dimensions  : 1760, 2640, 1  (nrow, ncol, nlayers)
> resolution  : 1, 1  (x, y)
> extent      : 0, 2640, -1759, 1  (xmin, xmax, ymin, ymax)
> projection  : +proj=utm +zone=31 +ellps=intl +units=m +no_defs
> values      : in memory
> min value   : 0
> max value   : 1
> So no NAs
> But:
>> summary(v)
>  SGRGBWBPS125F40_1 SGRGBWBPS125F40_2 SGRGBWBPS125F40_3
>  Min.   :18444     Min.   :  119.4   Min.   :    3
>  1st Qu.:33227     1st Qu.:25516.4   1st Qu.:15633
>  Median :37752     Median :36264.1   Median :22374
>  Mean   :40208     Mean   :36275.1   Mean   :29737
>  3rd Qu.:46899     3rd Qu.:48052.5   3rd Qu.:48116
>  Max.   :64992     Max.   :63667.1   Max.   :63826
>  NA's   :    5     NA's   :    1.0
>
>
> Any explanation to this? Not a big problem (can use na.rm) but I'm
> concerned on this problem implying
> the means not being correct.
>
> Agus
>