[R-sig-Geo] Problem with raster::extract

[Robert J. Hijmans]

This was caused by raster incorrectly interpreting the highest
possible INT2U value (65530) as NA. Now fixed (version 1.8-16), at
least for this case. Thanks for reporting, Robert

On Wed, Apr 27, 2011 at 8:52 AM, Robert J. Hijmans <r.hijmans at gmail.com> wrote:
> Hello Agus,
>
> The NaN suggests that there are only NA values for that polygon/layer.
> To investigate this I would do
>
> v <- extract(SGRGBF40, calibf2, nl=3)
>
> # i.e. do not use a function and then inspect
>
> v[[7]]
>
>
> All functions must accept an na.rm argument (even if they choose to
> ignore it. So for length you can do something like:
>
> vl <- extract(SGRGBF40, calibf2, fun=function(x,...)length(x), nl=3)
>
>
> Hope this helps, Robert
>
>> v
> [1] 387.8158 329.3913
>> v <- extract(r, polys, length)
> Error in FUN(X[[1L]], ...) :
>  2 arguments passed to 'length' which requires 1
>> v <- extract(r, polys, function(x,...)length(x))
>> v
> [1] 38 23
>>
>
>
> On Wed, Apr 27, 2011 at 3:14 AM, Agustin Lobo <alobolistas at gmail.com> wrote:
>> Hi!
>>
>> I'm doing
>>
>> 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)
>> #I make sure vector calibf2 and raster SGRGBF40 do overlap:
>> plot(subset(SGRGBF40,1))
>> plot(calibf2,add=T)
>>
>> #Calculate mean values for each polygon:
>> v <- extract(SGRGBF40, calibf2,fun=mean,na.rm=T,nl=3)
>> summary(v)
>> SGRGBF40data = cbind(calibf2 at data,v)
>>
>> But then, despite na.rm=T:
>>> SGRGBF40data[7,]
>>  ID code codetxt SGRGBWBPS125F40_1 SGRGBWBPS125F40_2 SGRGBWBPS125F40_3
>> 6  7    7      C7               NaN          39781.02                 3
>>
>>> summary(v)
>>  SGRGBWBPS125F40_1 SGRGBWBPS125F40_2 SGRGBWBPS125F40_3
>>  Min.   :18444     Min.   :  119.4   Min.   :    3
>>  1st Qu.:34098     1st Qu.:25875.7   1st Qu.:15633
>>  Median :37776     Median :36668.2   Median :22374
>>  Mean   :41314     Mean   :36882.2   Mean   :29737
>>  3rd Qu.:52236     3rd Qu.:49452.6   3rd Qu.:48116
>>  Max.   :65396     Max.   :65417.5   Max.   :63826
>>  NA's   :    3
>>
>> While there are no NA in the raster brick:
>>> summary(SGRGBF40)
>> Cells:  4646400
>> NAs  :  0 0 0
>>
>>            1     2     3
>> Min.     1777     3     3
>> 1st Qu. 34910 22460 18640
>> Median  42660 27540 24420
>> Mean    41070 28570 25170
>> 3rd Qu. 48470 33320 29060
>> Max.    65520 65530 64220
>>
>>
>> A possibility would be a division by 0 (weird, because all polygons
>> are large), so I've tried:
>>> vl <- extract(SGRGBF40, calibf2,fun=length,nl=3)
>> but...
>> Error in t(sapply(res[!i], function(x) apply(x, 2, FUN = fun, na.rm =
>> na.rm))) :
>>  error in evaluating the argument 'x' in selecting a method for
>> function 't': Error in FUN(newX[, i], ...) :
>>  2 arguments passed to 'length' which requires 1
>> Calls: sapply -> lapply -> FUN -> apply
>> Calls: extract -> extract -> .polygonValues -> t
>>
>> Any hint?
>>
>> Data:
>> http://dl.dropbox.com/u/3180464/SGRGBWBPS125F40.tif
>> http://dl.dropbox.com/u/3180464/SGRGBWBPS125F40.tfw
>> http://dl.dropbox.com/u/3180464/calibf2.zip
>>
>>> sessionInfo()
>> R version 2.13.0 (2011-04-13)
>> Platform: x86_64-pc-linux-gnu (64-bit)
>>
>> locale:
>>  [1] LC_CTYPE=en_US.UTF-8          LC_NUMERIC=C
>>  [3] LC_TIME=en_US.UTF-8           LC_COLLATE=en_US.UTF-8
>>  [5] LC_MONETARY=en_US.UTF-8       LC_MESSAGES=en_US.UTF-8
>>  [7] LC_PAPER=en_US.UTF-8          LC_NAME=en_US.UTF-8
>>  [9] LC_ADDRESS=en_US.UTF-8        LC_TELEPHONE=en_US.UTF-8
>> [11] LC_MEASUREMENT=en_US.UTF-8    LC_IDENTIFICATION=en_US.UTF-8
>>
>> attached base packages:
>> [1] grid      stats     graphics  grDevices utils     datasets  methods
>> [8] base
>>
>> other attached packages:
>> [1] gridExtra_0.7 ggplot2_0.8.9 rgdal_0.6-33  raster_1.7-29 sp_0.9-76
>> [6] reshape_0.8.4 plyr_1.4      proto_0.3-8   rkward_0.5.6
>>
>> loaded via a namespace (and not attached):
>> [1] lattice_0.19-26 tools_2.13.0
>>>
>>
>> Agus
>>
>