[R-sig-Geo] Combine two polygons

Fri Apr 28 17:41:48 CEST 2023

Hi Milu,

If you want to combine just a few of the thousands of features in
world then you will have to manage them individually.  Essentially
extract the subset to be combined, do the combination steps, and then
add them back to the world.  I am very not familiar with GADM and I
only have version 4.1, but all of the attributes appear to be
character type, which raises the question about what you want
aggregate() to do - it can't take the mean of a character attribute.
In the example below I show how you might take just the first value of
each attribute.  While this combining is possible, it doesn't mean
that you will have "correct" attributes with your new single feature.
I guess you'll have to figure out what you really want to do with
that.

### start here
library(sf)
library(dplyr)
# a function called by aggregate() for each attribute
first_item = function(x){
  x[[1]]
}
# read in the database downlaoded from https://gadm.org/data.html
world = read_sf("~/Downloads/gadm_410.gpkg")
# find the Cypriot features (we'll use this twice)
ix = grepl("Cyp", world$NAME_0, fixed = TRUE)
# pull them out of world
Cyp_features = dplyr::filter(world, ix)
# aggregate - which works for the geometry but is wrong in terms of
handling the attributes
Cyp_union = aggregate(Cyp_features,
                      sf::st_union(Cyp_features),
                      first_item)
# bind the world (less Cypriot features) with the combined Cypriot aggregation
new_world = dplyr::bind_rows(dplyr::filter(world, !ix), Cyp_union)
### end here

Does that help?

Ben

On Fri, Apr 28, 2023 at 8:28 AM Roger Bivand <Roger.Bivand using nhh.no> wrote:
>
> Briefly, is sf::aggregate() what you are looking for? Or terra::aggregate() if you read with terra::vect()?
>
> ---
> Roger Bivand
> Emeritus Professor
> Department of Economics
> Norwegian School of Economics, Bergen, Norway
>
> ________________________________
> Fra: Miluji Sb <milujisb using gmail.com>
> Sendt: fredag 28. april 2023, 14:18
> Til: Roger Bivand <Roger.Bivand using nhh.no>
> Kopi: Michael Sumner <mdsumner using gmail.com>; R-sig-geo mailing list <r-sig-geo using r-project.org>
> Emne: Re: [R-sig-Geo] Combine two polygons
>
> Thank you for your replies. I think I am misunderstanding something. I have read the shapefile using sf. I would like to combine polygons for Cyprus and Northern Cyprus only within the world shapefile and keep all the other polygons as they are.
>
> world <- sf::st_read(dsn = "~/gadm36_levels_shp", layer = "gadm36_0")
>
> Is something like the following correct? I was thinking of generating a group variable which is unique to all the countries except Cyprus and Northern Cyprus. I a probably making this too complicated.
>
> world %>%
>   group_by(group) %>%
>   summarise(geometry = sf::st_union(geometry)) %>%
>   ungroup()
>
> Thank you again.
>
> Best regards,
>
> Milu
>
> On Fri, Apr 28, 2023 at 8:56 AM Roger Bivand <Roger.Bivand using nhh.no<mailto:Roger.Bivand using nhh.no>> wrote:
> On Fri, 28 Apr 2023, Michael Sumner wrote:
>
> > fwiw, a method with geodata/terra
> >
> > cyp <- geodata::gadm(country = c("CYP", "XNC"), path = tempdir(), version =
> > "3.6")
> >
> > terra::aggregate(cyp)
> >
>
> and as per
> https://github.com/r-spatial/evolution/blob/main/pkgapi_230305_refs.csv.
>
> > see ?terra::writeVector to write it out
> >
> > (rgdal is very old and you should abandon it, it will be removed from
> > support by the end of this year)
>
> Thank you! Yes, sp will by default switch to using sf in place of rgdal
> for creating "CRS" objects, and under sp::spTransform in June, and all
> three retiring packages rgdal, rgeos, and maptools will be gone by October
> 2023. An sp release will appear in May with a transition guide.
>
> It seems that terra does not have an object corresponding to sp's "CRS" or
> sf's "crs". As far as I can see, one needs to create an empty
> SpatVector/SpatRaster object and assign a CRS, coerce to "Spatial", and
> extract its slot. If this makes sense, sp could use terra (and
> raster, needed for coercion) rather than sf where workflows did not
> otherwise use sf. Very grateful for input here or
> https://github.com/r-spatial/evolution/issues.
>
> Roger
>
> >
> >
> > (just as an aside, we can't hit the geodata vector urls directly with gdal
> > as we can with the rasters, as they are in an R specific format)
> >
> > Cheers, Mike
> >
> >
> > On Fri, Apr 28, 2023 at 10:52 AM Ben Tupper <btupper using bigelow.org<mailto:btupper using bigelow.org>> wrote:
> >
> >> Hi,
> >>
> >> You might try st_union() from the sf package
> >> https://r-spatial.github.io/sf/articles/sf3.html#geometrical-operations
> >>
> >>
> >>
> >> On Thu, Apr 27, 2023 at 5:37 PM Miluji Sb <milujisb using gmail.com<mailto:milujisb using gmail.com>> wrote:
> >>
> >>> Dear all,
> >>>
> >>> I am using the country-level shapefile from GADM. I would like to merge
> >>> Cyprus and North Cyprus into one polygon. Is this possible? I am using
> >>> rgdal to read the shapefile, subset, and then rewrite it. Any help will
> >> be
> >>> highly appreciated.
> >>>
> >>> ##
> >>> world <- readOGR("~/gadm36_levels_shp", layer = "gadm36_0")
> >>>
> >>> Best regards,
> >>>
> >>> Milu
> >>>
> >>>         [[alternative HTML version deleted]]
> >>>
> >>> _______________________________________________
> >>> R-sig-Geo mailing list
> >>> R-sig-Geo using r-project.org<mailto:R-sig-Geo using r-project.org>
> >>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> >>>
> >>
> >>         [[alternative HTML version deleted]]
> >>
> >> _______________________________________________
> >> R-sig-Geo mailing list
> >> R-sig-Geo using r-project.org<mailto:R-sig-Geo using r-project.org>
> >> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> >>
> >
> >
> >
>
> --
> Roger Bivand
> Emeritus Professor
> Department of Economics, Norwegian School of Economics,
> Postboks 3490 Ytre Sandviken, 5045 Bergen, Norway.
> e-mail: Roger.Bivand using nhh.no<mailto:Roger.Bivand using nhh.no>
> https://orcid.org/0000-0003-2392-6140
> https://scholar.google.no/citations?user=AWeghB0AAAAJ&hl=en_______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo using r-project.org<mailto:R-sig-Geo using r-project.org>
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>
>
>         [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo using r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo