[R-sig-Geo] sp::overlay() not conserving row.names ?

[Agustin Lobo]

Yes!

I'm doing:
> delme2 <- over(q1km, rndp,fn = mean)

And now not only the row.names are correct and the order of the 2
arguments makes more sense, but the dimensions of the result are also
the same
that that of the first argument
(which implies rows of NA for those polygons with no points: the user
can get rid of
these rows with na.omit() later, for me having the NA is a lot better):

> delme2[1:3,]
        ID      val
0 356.3333 760.5662
1 454.3333 541.5580
2 305.0000 467.0864
> q1km at data[1:3,]
  ID   XMIN   XMAX   YMIN   YMAX
0  0 -8.994 -8.594 43.595 43.995
1  1 -8.594 -8.194 43.595 43.995
2  2 -8.194 -7.794 43.595 43.995

> dim(delme2)
[1] 299   3
> dim(q1km at data)
[1] 299   6

I can just join this table to the one of q1km:
> delme2 = cbind(idname=row.names(delme2),delme2)
> delme2[1:3,]
  idname       ID      val
0      0 356.3333 760.5662
1      1 454.3333 541.5580
2      2 305.0000 467.0864

> q1kmv2 = q1km
> q1kmv2 <- mijoin(q1kmv2,delme2[,-2], by.x=1,by.y=1)
> q1kmv2 at data[1:3,]
  ID   XMIN   XMAX   YMIN   YMAX idname      val
0  0 -8.994 -8.594 43.595 43.995      0 760.5662
1  1 -8.594 -8.194 43.595 43.995      1 541.5580
2  2 -8.194 -7.794 43.595 43.995      2 467.0864
3  3 -7.794 -7.394 43.595 43.995      3 487.6461

Thanks!
Agus

2011/7/6 Edzer Pebesma <edzer.pebesma at uni-muenster.de>:
> Agus, sp::overlay will be deprecated at some stage in favour of
> sp::over, which is a more consistent and complete approach to the same
> problem. For instance, overlay(x,y) would do the same as overlay(y,x),
> which is not good, if you think about it a bit longer (my own mistake,
> long time ago).
>
> Could you please check if sp::over has the same problems? See also
> http://cran.r-project.org/web/packages/sp/vignettes/over.pdf
>
> On 07/06/2011 05:28 PM, Agustin Lobo wrote:
>> I'm using sp::overlay between a SpPolDF and a SpPointsDF:
>>
>>> delme <- overlay(rndp,q1km, fn = mean)
>>> class(q1km)
>> [1] "SpatialPolygonsDataFrame"
>> attr(,"package")
>> [1] "sp"
>>> class(rndp)
>> [1] "SpatialPointsDataFrame"
>> attr(,"package")
>> [1] "sp"
>>> class(delme)
>> [1] "data.frame"
>>
>> but I'm confused by the fact that the row.names(delme) are not the same
>> as the row.names(q1km):
>>> delme[1:10,]
>>           ID       id      val
>> X1  356.3333 357.3333 643.3162
>> X2  454.3333 455.3333 653.1833
>> X3  305.0000 306.0000 101.8930
>> X4  486.7143 487.7143 448.4612
>> X5  456.2500 457.2500 544.1646
>> X6  255.4000 256.4000 139.7507
>> X7  403.6667 404.6667 659.8114
>> X8  462.5000 463.5000 537.7209
>> X9  629.2000 630.2000 399.1040
>> X10 416.5000 417.5000 278.2879
>>> q1km at data[1:10,]
>>    idname ID   XMIN   XMAX   YMIN   YMAX
>> X0     X0  0 -8.994 -8.594 43.595 43.995
>> X1     X1  1 -8.594 -8.194 43.595 43.995
>> X2     X2  2 -8.194 -7.794 43.595 43.995
>> X3     X3  3 -7.794 -7.394 43.595 43.995
>> X4     X4  4 -7.394 -6.994 43.595 43.995
>> X5     X5  5 -6.994 -6.594 43.595 43.995
>> X6     X6  6 -6.594 -6.194 43.595 43.995
>> X7     X7  7 -6.194 -5.794 43.595 43.995
>> X8     X8  8 -5.794 -5.394 43.595 43.995
>> X9     X9  9 -5.394 -4.994 43.595 43.995
>>
>> Is this the way it has to be? If it is, then it's kind of inconvenient
>> as I would like to join the new table "delme" to q1km at data.
>> It seems I just have to subtract 1 to the numeric part of
>> row.names(delme), but would like to make sure.
>>
>> Agus
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>
> --
> Edzer Pebesma
> Institute for Geoinformatics (ifgi), University of Münster
> Weseler Straße 253, 48151 Münster, Germany. Phone: +49 251
> 8333081, Fax: +49 251 8339763  http://ifgi.uni-muenster.de
> http://www.52north.org/geostatistics      e.pebesma at wwu.de
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>