Creating spatialpolygons dataframe from list of polygons

Question

I am currently trying to create a polygon shapefile from a list of polygons (study areas for biodiversity research).

Currently these polygons are stored in a list in this format:

$SEW22
     [,1]    [,2]
[1,] 427260.4 5879458
[2,] 427161.4 5879472
[3,] 427175.0 5879571
[4,] 427273.9 5879557
[5,] 427260.4 5879458

$SEW23
     [,1]    [,2]
 [1,] 418011.0 5867216
 [2,] 417912.0 5867230
 [3,] 417925.5 5867329
 [4,] 418024.5 5867315
 [5,] 418011.0 5867216

I tried to simply write them as shpfile with writeOGR but the following error occurs:

> #write polygons to shp
> filenameshp <- paste('Forestplots')
> layername <- paste('Forestplots')
> writeOGR(obj=forest, dsn = filenameshp, 
+          layer=layername, driver="ESRI Shapefile", overwrite_layer =     TRUE)
Error in writeOGR(obj = forest, dsn = filenameshp, layer = layername,  : 
 inherits(obj, "Spatial") is not TRUE

I read this tutorial by Barry Rowlingson to create spatialpolygons and thought I should probably first create a dataframe and did this:

forestm<-do.call(rbind,forest)

but this returned nothing useful as you can imagine, plus it lost the names of the plots.

As I am still new to R I also tried lots of different other approaches which sensefulness I could not fully judge but none returned what I hoped for and so I spare you with these random approaches.....

I am looking forward to your propositions.

Many thanks

P.S. I also tried the following as described in the spatialpolygons{sp} package:

> Polygons(forest, ID)
Error in Polygons(forest, ID) : srl not a list of Polygon objects

Weihuang Wong · Accepted Answer · 2018-07-05T14:43:58.837

You can follow the approach described in this answer: https://gis.stackexchange.com/questions/18311/instantiating-spatial-polygon-without-using-a-shapefile-in-r.

Here's how to apply the approach to your case. First, I create a list of matrices as in your sample data:

forest <- list(
  "SEW22" = matrix(c(427260.4, 5879458, 427161.4, 5879472, 427175.0, 5879571, 427273.9, 5879557, 427260.4, 5879458),
                   nc = 2, byrow = TRUE),
  "SEW23" = matrix(c(418011.0, 5867216, 417912.0, 5867230, 417925.5, 5867329, 418024.5, 5867315, 418011.0, 5867216),
                   nc = 2, byrow = TRUE)
  )

Now

library(sp)
p <- lapply(forest, Polygon)
ps <- lapply(seq_along(p), function(i) Polygons(list(p[[i]]), ID = names(p)[i]))
sps <- SpatialPolygons(ps)
sps_df <- SpatialPolygonsDataFrame(sps, data.frame(x = rep(NA, length(p)), row.names = names(p)))

In the first step, we iterate through the list of matrices and apply the Polygon function to each matrix to create a list of Polygon objects. In the second step, we iterate through this list to create a Polygons object, setting the ID of each element in this object to the corresponding name in the original list (e.g. "SEW22", "SEW23"). The third step creates a SpatialPolygons object. Finally, we create a SpatialPolygonsDataFrame object. Here I have a dummy dataframe populated with NAs (note that the row names must correspond to the polygon IDs).

Finally, write the data

rgdal::writeOGR(obj = sps_df,
                dsn = "Forestplots",
                layer = "Forestplots",
                driver = "ESRI Shapefile",
                overwrite_layer = TRUE)

This creates a new folder in your working directory:

list.files()
# [1] "Forestplots"
list.files("Forestplots")
# [1] "Forestplots.dbf" "Forestplots.shp" "Forestplots.shx"

Consult the linked answer for more details.

Thank you very much for this complete answer! Is there a way to create the matrices automatically? I have 150 plots and assuming I would be able to write all coordinates of all vertices of each polygon correctly in code, it would take a lot of time.... — S.K., Jul 05 '18 at 13:06
Welcome to Stack Overflow! If this answer resolved your question, please mark it as accepted. If you have follow-up questions, please open a new question -- the comments section is not designed for answers. — Weihuang Wong, Jul 05 '18 at 13:55
I just saw your other question. I think you may have misunderstood my answer. The `forest <- list(...)` part of the answer was just so I could re-create your original list of matrices. Obviously you don't have to do this if you already have the `forest` object. — Weihuang Wong, Jul 05 '18 at 14:32
Thank you very much Weihuang, actually I was just in the process of learning lapply function for my second question. Works like a treat. Many thanks! — S.K., Jul 05 '18 at 14:41
I realised, that using the writeOGR function above, the plot names from the spatialpolygonsdataframe get lost. Is there a way to maintain the plotnames? — S.K., Jul 06 '18 at 08:38
Please submit a new question with a minimal replicable example. — Weihuang Wong, Jul 06 '18 at 10:59

Creating spatialpolygons dataframe from list of polygons

1 Answers1