Skip to content

how to create several polygons from a list of vertices

7 messages · Vijay Lulla, obrl soil, Michael Sumner +1 more

Original
Antonio Silva
# 
Hi,

I have a data.frame with the vertices (lon / lat) and codes from several
squares (more than 500 in the real dataset).
I want to create an object with these polygons (squares) and after this
export it as a shapefile.
With the script below I can draw one square.
library(sp)
P1 = Polygon(vertices[1:4,1:2])
Ps1 = SpatialPolygons(list(Polygons(list(P1), ID = "1")),
proj4string=CRS("+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs"))
plot(Ps1, axes = TRUE)

Now I'm trying to create one object with all squares at once.
Is it possible?

Thanks a lot,

Ant?nio Olinto

sample data:vertices
   lon lat cod
1  -33 -23   1
2  -32 -23   1
3  -32 -22   1
4  -33 -22   1
5  -32 -23   2
6  -31 -23   2
7  -31 -22   2
8  -32 -22   2
9  -31 -23   3
10 -30 -23   3
11 -30 -22   3
12 -31 -22   3
13 -33 -22   4
14 -32 -22   4
15 -32 -21   4
16 -33 -21   4
17 -32 -22   5
18 -31 -22   5
19 -31 -21   5
20 -32 -21   5
21 -31 -22   6
22 -30 -22   6
23 -30 -21   6
24 -31 -21   6
Vijay Lulla
# 
Maybe something like this?

poly <- SpatialPolygons(list(Polygons(tapply(seq_len(nrow(vertices)),
                                             vertices$cod,
                                             function(x)
Polygon(vertices[x,1:2])), ID="1")),
                        proj4string=CRS("+proj=longlat +ellps=WGS84
+datum=WGS84 +no_defs"))
On Tue, Aug 14, 2018 at 4:17 PM Antonio Silva <aolinto.lst at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Antonio Silva
    

    

      
      #
    

  


  

      
Thanks Lulla,

Nice solution. I could also export it as a shapefile after transforming it
to a spatial polygon dataframe.

The problem is that I could not "individualize" the squares in a multipart
layer. They all have the same ID. I tried to change this without success:
"Single ID required".

The attribute table of the shapefile should have 6 lines in my example and
not only one.

Any other option?

Thanks again,

Antonio Olinto


Em ter, 14 de ago de 2018 ?s 18:10, Vijay Lulla <vijaylulla at gmail.com>
escreveu:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      obrl soil
    

    

      
      #
    

  


  

      
Hi Antonio,

have you tried with sf? Like:

library(sf)

pts <-
  tibble::tribble(~ID,  ~x,  ~y, ~grp,
                   1 , -33, -23,    1,
                   2 , -32, -23,    1,
                   3 , -32, -22,    1,
                   4 , -33, -22,    1,
                   5 , -32, -23,    2,
                   6 , -31, -23,    2,
                   7 , -31, -22,    2,
                   8 , -32, -22,    2,
                   9 , -31, -23,    3,
                   10, -30, -23,    3,
                   11, -30, -22,    3,
                   12, -31, -22,    3,
                   13, -33, -22,    4,
                   14, -32, -22,    4,
                   15, -32, -21,    4,
                   16, -33, -21,    4,
                   17, -32, -22,    5,
                   18, -31, -22,    5,
                   19, -31, -21,    5,
                   20, -32, -21,    5,
                   21, -31, -22,    6,
                   22, -30, -22,    6,
                   23, -30, -21,    6,
                   24, -31, -21,    6)

squares <- split(pts, pts$grp)
squares <- lapply(squares, function(g) {
  # get just coords
  g <- as.matrix(g[, c(2,3)])
  # repeat first point last to close poly
  g <- rbind(g, g[1, ])
  # convert to an sf polygon object
  gp <- sf::st_polygon(list(g))
  # make sure the vertices are in an order
  # that complies with the simple
  # features standard
  gp <- sf::st_buffer(gp, 0L)
})
# turn list of polygons into geometry column
squares_sfc <- sf::st_sfc(squares) # can add crs = ?? to this call
# add an ID to make an sf data frame
squares_sf <- sf::st_sf('ID' = seq(6), 'geometry' = squares_sfc)

# if you still need to use sp for whatever reason
squares_sp <- as(squares_sf, "Spatial")

Regards,
@obrl_soil

        
On Wed, Aug 15, 2018 at 8:03 AM, Antonio Silva <aolinto.lst at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Michael Sumner
    

    

      
      #
    

  


  

      
Here's another way with spbabel.


library(dplyr)
library(spbabel)
pts <-
  tibble::tribble(~ID,  ~x,  ~y, ~grp,
                   1 , -33, -23,    1,
                   2 , -32, -23,    1,
                   3 , -32, -22,    1,
                   4 , -33, -22,    1,
                   5 , -32, -23,    2,
                   6 , -31, -23,    2,
                   7 , -31, -22,    2,
                   8 , -32, -22,    2,
                   9 , -31, -23,    3,
                   10, -30, -23,    3,
                   11, -30, -22,    3,
                   12, -31, -22,    3,
                   13, -33, -22,    4,
                   14, -32, -22,    4,
                   15, -32, -21,    4,
                   16, -33, -21,    4,
                   17, -32, -22,    5,
                   18, -31, -22,    5,
                   19, -31, -21,    5,
                   20, -32, -21,    5,
                   21, -31, -22,    6,
                   22, -30, -22,    6,
                   23, -30, -21,    6,
                   24, -31, -21,    6)


## objects and branches (parts) are the same level

## objects and branches (parts) are the same level
## and we maintain "grp" as the object attribute data
x <- pts %>% transmute(grp = grp, x_ = x, y_ = y,
               object_ = grp, branch_ = grp,
               order_ = ID, island_ = TRUE) %>% spbabel::sp()

        
On Wed, 15 Aug 2018 at 09:25 obrl soil <obrlsoilau at gmail.com> wrote:

        

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Vijay Lulla
    

    

      
      #
    

  


  

      
Maybe you can try this then?

polys <- SpatialPolygons(lapply(unique(vertices$cod),
                         function(x) {
                           Polygons(list(Polygon(vertices[vertices$cod ==
x, 1:2])), ID=x)
                         } ))
HTH,
Vijay.

        
On Tue, Aug 14, 2018 at 6:03 PM Antonio Silva <aolinto.lst at gmail.com> wrote:

        

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Antonio Silva
    

    

      
      #
    

  


  

      
Hi Vijay, Michael and Obrl

Thanks for the answers, finally I could do the job.

I learned a lot too!

Best regards

Antonio Olinto

Em Ter, 14 de ago de 2018 11:19 PM, Vijay Lulla <vijaylulla at gmail.com>
escreveu: