如何删除 0 个或多个标记（单词），它们可能会按顺序出现或中间有数据出现？

Question

How to extract only the country names from a variable such as the following.

tibble::tribble(
    ~country, 
    '{"United States"}', 
    '{NULL}', 
    '{NULL,NULL}', 
    '{"United States",NULL,Netherlands}', 
    '{Germany}', 
    '{Canada}', 
    '{NULL,NULL}', 
    '{Chile,"United States"}', 
    '{NULL,NULL,NULL}', 
    '{NULL,China, NULL}', 
)

• NULL can come up sequentially or not and up to 15 time in a single observation.

• Countries with multiple words, such as "United States" come up quoted, otherwise they are all unquoted.

It is somewhat easy to do in multiple runs, such as removing all NULL s, then removing the duplicated commas, and then the parenthesis, but I was aiming for a more efficient way of achieving something towards the following:

tibble::tribble(
    ~country, 
    'United States', 
    NA, 
    NA, 
    'United States,Netherlands', 
    'Germany', 
    'Canada', 
    NA, 
    'Chile,United States', 
    NA, 
    'China', 
)

Answer 1

A bit brute-force with gsub s, but it works.

dat$out <- gsub("^,|,$", "",
                trimws(gsub('NULL,?|["{}]', '', dat$country)))
dat
# # A tibble: 10 x 2
#    country                                out                        
#    <chr>                                  <chr>                      
#  1 "{\"United States\"}"                  "United States"            
#  2 "{NULL}"                               ""                         
#  3 "{NULL,NULL}"                          ""                         
#  4 "{\"United States\",NULL,Netherlands}" "United States,Netherlands"
#  5 "{Germany}"                            "Germany"                  
#  6 "{Canada}"                             "Canada"                   
#  7 "{NULL,NULL}"                          ""                         
#  8 "{Chile,\"United States\"}"            "Chile,United States"      
#  9 "{NULL,NULL,NULL}"                     ""                         
# 10 "{NULL,China, NULL}"                   "China"                    

From here, you can replace the empty strings with "" with

dat$out[!nzchar(dat$out)] <- NA