Snippet to show how to retrieve and add scientific names to a dataframe containing a list of non unique vernacular names based on GBIF Backbone. The most likely match only is returned. If the vernacular name cannot be matched, NA is returned

get_scientific_names_from_vernacular_names.R

#' load packages
library(rgbif)
library(tidyverse)

#' example input
vernacular_names_df <- tibble(
  id = c(1,2,3,4,5,6,7,8),
  vernacular_name = c("Beenvissen",
                      "Bruine beer",
                      "Bont zandoogje",
                      "Bont zandoogje",
                      "Bruine beer",
                      "Beenvissen",
                      "Muscusrat > 400g",
                      "doodaars")
)

#' define get_vernacular_name() core function
#' input: a vernacular name
#' output: the best matched scientific name from the GBIF Backbone, NA_character_ if no match found
get_vernacular_name <- function(vn) {
  names <-
    name_lookup(vn,
                datasetKey = "d7dddbf4-2cf0-4f39-9b2a-bb099caae36c",
                limit = 1)$data # this returns the most likely taxon
  if (nrow(names) > 0) {
    names$scientificName
  } else {
    NA_character_
  }
}

n_vernacular_names_df <-
  vernacular_names_df %>%
  
  # group  by vernacular name and compact the data
  group_by(vernacular_name) %>%
  nest() %>%
  
  # find scientific name for each (distinct) vernacular name
  mutate(scientificName = map_chr(vernacular_name,get_vernacular_name)) %>%
  
  # ungroup result
  ungroup() %>%
  
  # remove unneeded columns
  select(-one_of("data")) %>%
  
  # add other columns from input df vernacular_names_df
  right_join(vernacular_names_df, by = "vernacular_name") %>%
  
  # set new column scientificName at the right side
  select(all_of(names(vernacular_names_df)), scientificName) %>%
  
  # reorder rows based on original order in input df
  right_join(vernacular_names_df,
             by = names(vernacular_names_df))

#' show results
n_vernacular_names_df

Another interesting remark from @fredericpiesschaert:

één van de problemen is dat er op alle vernacular names onafhankelijk van de taal gecheckt wordt, waardoor je bv bij hop een hoop taxa krijgt die niets te maken hebben met wat we willen krijgen. Als je de taal als parameter kan meegeven zou dat dus ook al veel schelen denk ik