1

I'm trying to intersect two sf objects for the US (one at the township level and the other one at the census tract level). I'm getting both using tigris and tidycensys. My final goal is to have an unique sf object with information at the township level (with information from both the originals township and census tract level sf objects). And after I do this intersection, I want to export this sf object using st_write from the sf package. Here is the code I've used:

library(tigris)
library(sf)
library(purrr)
library(tidycensus)
library(tidyr)
library(dplyr)

##Data at township level 
#---------------------------#
MN_Township_SHP <- county_subdivisions("Minnesota", cb = TRUE)%>% st_transform(., crs=32618)
MN_Township_SHP$County <- substr(MN_Township_SHP$NAMELSADCO,1,nchar(MN_Township_SHP$NAMELSADCO)-7)
Dataset <- MN_Township_SHP

#Data at census tract level
#---------------------------#
Sys.getenv("CENSUS_API_KEY") 

my_vars <- 
  c(total_pop        = "B01003_001",
    race_denominator = "B02001_001", #Total
    white            = "B02001_002")
mn <- unique(fips_codes$state)[24]
MN_CensusTract_SHP <-map_df(mn, function(x) {
  get_acs(geography = "tract", 
          geometry = T, 
          variables = my_vars, 
          state = mn) })
MN_CensusTract_SHP <- MN_CensusTract_SHP %>% dplyr::select(-moe) 
Social_Dat <-
  MN_CensusTract_SHP %>%
  as.data.frame() %>%
  pivot_wider(names_from = variable,
              values_from = c(estimate)) %>%
  dplyr::mutate(year=2021) %>% 
  dplyr::rename_all(funs(paste0("ACS_", .))) 
Social_Dat$ACS_year <- as.double(Social_Dat$ACS_year)
Social_Dat$ACS_GEOID <- as.double(Social_Dat$ACS_GEOID)
Social_Dat <- st_as_sf(Social_Dat, sf_column_name = 'ACS_geometry')%>% st_transform(., crs=32618)

#Intersection between township and census tract levels
#---------------------------#
final_df <- st_intersection(Dataset, Social_Dat, all=TRUE)

#Export sf object as shapefile 
#---------------------------#
st_write(final_df, "Input_Intermediate/final_df.shp",  delete_layer = TRUE)

However, when I run this final step, I get the following error:

"Error in CPL_write_ogr(obj, dsn, layer, driver, as.character(dataset_options), : Write error"

Does anyone know how to solve this? I've tried so many ways that I found in google, but none of them worked for me. Many thanks in advance!!!

Adriana Castillo Castillo
  • 170
  • 1
  • 9

2 Answers2

2

You can use the function st_collection_extract and get only the polygon type from your geometry column, and then you can proceed and use st_write again.

df <- st_collection_extract(final_df, type = "POLYGON")
st_write(df, "Input_Intermediate/final_df.shp",  delete_layer = TRUE)
rodrigo franco
  • 161
  • 5
  • Hi Rodrigo, many thanks for your answer. I have been struggling with this issue with other databases, I tried your solution, but it didn't work out for the other cases. Do you have any idea of how can I solve this in a more general way? Many thanks in advance! – Adriana Castillo Castillo Mar 08 '23 at 14:00
0

I was able to solve this issue by doing the following:

  1. Check possible variable names that may cause problems. I.e. I started only with the first 2 columns and then the following. One of them was giving me problems, so I changed the name.
  2. Make sure the file name you are using to save your shapefile is short. In my case the error was because both the path+filename were too long.
Adriana Castillo Castillo
  • 170
  • 1
  • 9