compute_panel full_join fails, but not inner_join nor anti_join

[EvaMaeRey]

I've been using inner_join in #10 and now #54's second geodata packages example. This means that regions will only be represented if they are in the plot data and the reference data. And I've kind of tried to convince myself that this is what you want, but maybe that is motivated and mostly because of this full_join error that is popping up. Interestingly, anti_join which gives you the reference data regions, but not the plot data regions, works!

library(tidyverse)

nc_ref <- sf::st_read(system.file("shape/nc.shp", package="sf")) |>
  select(county_name = NAME, fips = FIPS) %>% 
    ggplot2::StatSf$compute_panel(coord = ggplot2::CoordSf) |>
    ggplot2::StatSfCoordinates$compute_group(coord = ggplot2::CoordSf)# sf object preserves geometry column on save

nc_ref |> head()
#> Simple feature collection with 6 features and 8 fields
#> Geometry type: MULTIPOLYGON
#> Dimension:     XY
#> Bounding box:  xmin: -81.74107 ymin: 36.07282 xmax: -75.77316 ymax: 36.58965
#> Geodetic CRS:  NAD27
#>   county_name  fips                       geometry      xmin      xmax     ymin
#> 1        Ashe 37009 MULTIPOLYGON (((-81.47276 3... -84.32385 -75.45698 33.88199
#> 2   Alleghany 37005 MULTIPOLYGON (((-81.23989 3... -84.32385 -75.45698 33.88199
#> 3       Surry 37171 MULTIPOLYGON (((-80.45634 3... -84.32385 -75.45698 33.88199
#> 4   Currituck 37053 MULTIPOLYGON (((-76.00897 3... -84.32385 -75.45698 33.88199
#> 5 Northampton 37131 MULTIPOLYGON (((-77.21767 3... -84.32385 -75.45698 33.88199
#> 6    Hertford 37091 MULTIPOLYGON (((-76.74506 3... -84.32385 -75.45698 33.88199
#>       ymax         x        y
#> 1 36.58965 -81.49496 36.42112
#> 2 36.58965 -81.13241 36.47396
#> 3 36.58965 -80.69280 36.38828
#> 4 36.58965 -75.93852 36.30697
#> 5 36.58965 -77.36988 36.35211
#> 6 36.58965 -77.04217 36.39709

inner join as compute

# layer_county w/ inner join
compute_panel_nc_county <- function(data, scales){

nc_ref  |>
    inner_join(data)

}

StatNCcounties <- ggproto("StatNCcounties",
                        Stat,
                        compute_panel = compute_panel_nc_county,
                        default_aes = aes(label = after_stat(county_name)),
                        required_aes = c("county_name|fips"))


tribble(~county, ~pop,
        "Mecklenburg", 3000) |>
ggplot() +
  aes(county_name = county) + 
  geom_sf(stat = StatNCcounties) + 
  aes(fill = pop)
#> Joining with `by = join_by(county_name)`

Full join compute fails

# layer_county w/ full join fails: Error in `scale_apply()`: `scale_id` must not contain any "NA".
compute_panel_nc_county_fulljoin <- function(data, scales){

nc_ref  |>
    full_join(data)

}

tribble(~county, ~pop,
        "Mecklenburg", 3000) |>
  rename(county_name = county) |>
  compute_panel_nc_county_fulljoin()
#> Joining with `by = join_by(county_name)`
#> Simple feature collection with 100 features and 9 fields
#> Geometry type: MULTIPOLYGON
#> Dimension:     XY
#> Bounding box:  xmin: -84.32385 ymin: 33.88199 xmax: -75.45698 ymax: 36.58965
#> Geodetic CRS:  NAD27
#> First 10 features:
#>    county_name  fips      xmin      xmax     ymin     ymax         x        y
#> 1         Ashe 37009 -84.32385 -75.45698 33.88199 36.58965 -81.49496 36.42112
#> 2    Alleghany 37005 -84.32385 -75.45698 33.88199 36.58965 -81.13241 36.47396
#> 3        Surry 37171 -84.32385 -75.45698 33.88199 36.58965 -80.69280 36.38828
#> 4    Currituck 37053 -84.32385 -75.45698 33.88199 36.58965 -75.93852 36.30697
#> 5  Northampton 37131 -84.32385 -75.45698 33.88199 36.58965 -77.36988 36.35211
#> 6     Hertford 37091 -84.32385 -75.45698 33.88199 36.58965 -77.04217 36.39709
#> 7       Camden 37029 -84.32385 -75.45698 33.88199 36.58965 -76.18290 36.36249
#> 8        Gates 37073 -84.32385 -75.45698 33.88199 36.58965 -76.72199 36.43576
#> 9       Warren 37185 -84.32385 -75.45698 33.88199 36.58965 -78.11342 36.42681
#> 10      Stokes 37169 -84.32385 -75.45698 33.88199 36.58965 -80.23459 36.40106
#>    pop                       geometry
#> 1   NA MULTIPOLYGON (((-81.47276 3...
#> 2   NA MULTIPOLYGON (((-81.23989 3...
#> 3   NA MULTIPOLYGON (((-80.45634 3...
#> 4   NA MULTIPOLYGON (((-76.00897 3...
#> 5   NA MULTIPOLYGON (((-77.21767 3...
#> 6   NA MULTIPOLYGON (((-76.74506 3...
#> 7   NA MULTIPOLYGON (((-76.00897 3...
#> 8   NA MULTIPOLYGON (((-76.56251 3...
#> 9   NA MULTIPOLYGON (((-78.30876 3...
#> 10  NA MULTIPOLYGON (((-80.02567 3...

StatNCcounties$compute_panel <- compute_panel_nc_county_fulljoin

tribble(~county, ~pop,
        "Mecklenburg", 3000) |>
ggplot() +
  aes(county_name = county) + 
  geom_sf(stat = StatNCcounties) + 
  aes(fill = pop)
#> Joining with `by = join_by(county_name)`
#> Error in `scale_apply()`:
#> ! `scale_id` must not contain any "NA".

Antijoin compute doesn't error

# layer_county w/ anti join fails: Error in `scale_apply()`: `scale_id` must not contain any "NA".
compute_panel_nc_county_antijoin <- function(data, scales){

nc_ref  |>
    anti_join(data)

}


StatNCcounties$compute_panel <- compute_panel_nc_county_antijoin


tribble(~county, ~pop,
        "Mecklenburg", 3000) |>
ggplot() +
  aes(county_name = county) + 
  geom_sf(stat = StatNCcounties) + 
  aes(fill = pop)
#> Joining with `by = join_by(county_name)`

^{Created on 2024-11-15 with reprex v2.1.0}