ropensci
"All inputs to rbind.fill must be data.frames" error on getting project's observations
I am trying grab project observations but I am getting an rbind error.
``` r
library(rinat)
project_slug <- "unt-herpetology-spr-2025"
inat_data <- get_inat_obs_project(project_slug, type="observations")
#> 3611 records
#> Getting records 0-200
#> Getting records up to 400
#> Getting records up to 600
#> Getting records up to 800
#> Getting records up to 1000
#> Getting records up to 1200
#> Getting records up to 1400
#> Getting records up to 1600
#> Getting records up to 1800
#> Getting records up to 2000
#> Getting records up to 2200
#> Getting records up to 2400
#> Getting records up to 2600
#> Getting records up to 2800
#> Getting records up to 3000
#> Getting records up to 3200
#> Getting records up to 3400
#> Getting records up to 3600
#> Getting records up to 3800
#> Done.
#> Error: All inputs to rbind.fill must be data.frames
Created on 2025-03-08 with reprex v2.1.1
I also noticed I can grab observations from some projects but not others. How does rinat interface with different project options? The above project has a member's only setting. The below (presumably) does not. Is it related to this issue? https://forum.inaturalist.org/t/select-observations-to-batch-download-from-list-of-observation-ids/61657
``` r
library(rinat)
project_slug <- "test"
inat_data <- get_inat_obs_project(project_slug, type="observations")
#> 3 records
#> Getting records 0-200
#> Done.
Created on 2025-03-08 with reprex v2.1.1
Indeed, I can reproduce the issue. It looks like a big mismatch between the API's reported "project_observations_count" (currently 4368) and what the website shows (currently 132). I don't know why there is such a large discrepancy.
I had previously added a step to cater for a mismatch, but it would only work if the last element of the list was empty. In this case, more than one element is empty. I will try to find time to release an update soon, but in the meantime you can use this modified version of the function, get_inat_obs_project2():
library(jsonlite)
library(httr)
library(plyr)
get_inat_obs_project2 <- function (grpid,
type = c("observations", "info"),
raw = FALSE) {
if (!curl::has_internet()) {
message("No Internet connection.")
return(invisible(NULL))
}
base_url <- "http://www.inaturalist.org/"
if (httr::http_error(base_url)) {
message("iNaturalist API is unavailable.")
return(invisible(NULL))
}
argstring <- switch(match.arg(type), observations = "obs",
info = "info")
url <- paste0(base_url, "projects/", grpid, ".json")
xx <- fromJSON(content(GET(url), as = "text"))
recs <- xx$project_observations_count
dat <- NULL
if (is.null(recs))
(return(dat))
message(paste(recs, "records\n"))
if (argstring == "info") {
output <- list()
output[["title"]] <- xx$title
output[["description"]] <- xx$description
output[["slug"]] <- xx$slug
output[["created_at"]] <- xx$created_at
output[["id"]] <- xx$id
output[["location"]] <- c(as.numeric(xx$lat), as.numeric(xx$long))
output[["place_id"]] <- xx$place_id
output[["taxa_number"]] <- xx$observed_taxa_count
output[["taxa_count"]] <- xx$project_observations_count
if (raw) {
output[["raw"]] <- xx
}
return(output)
}
else if (argstring == "obs") {
per_page <- 200
if (recs%%per_page == 0) {
loopval <- recs%/%per_page
}
if (recs >= 10000) {
warning("Number of observations in project greater than current API limit.\nReturning the first 10000.\n")
loopval <- 10000/per_page
}
else {
loopval <- (recs%/%per_page) + 1
}
obs_list <- vector("list", loopval)
for (i in 1:loopval) {
url1 <- paste0(base_url, "observations/project/",
grpid, ".json?page=", i, "&per_page=", per_page)
if (i == 1) {
message(paste0("Getting records 0-", per_page))
}
if (i > 1) {
message(paste0("Getting records up to ", i *
per_page))
}
obs_list[[i]] <- fromJSON(content(GET(url1), as = "text"),
flatten = TRUE)
}
message("Done.\n")
# remove empty elements
obs_list <- obs_list[lengths(obs_list) >= 1]
project_obs <- do.call("rbind.fill", obs_list)
if (recs != nrow(project_obs)) {
message("Note: mismatch between number of observations reported and returned by the API.")
}
return(project_obs)
}
}
This works for me, hopefully it gets you what you need.
