从列表子集以逗号分隔的字符串

Question

This seems like a straightforward operation, but I seem to be stuck and am looking for pointers.

I have a dataframe of authors and their associated publications. In the author column, there are often times multiple authors for a single article in a semicolon delimited list. Here's a small subset:

structure(list(author = c("Moscatelli, Adriana; Nishina, Adrienne", 
"Asangba, Abigail", "Stewart, Abigail", "Redmond-Sanogo, Adrienne; Lee, Ahlam", 
"Purnamasari, Agustina; Lee, Ahlam; Moscatelli, Adriana", 
"Nishina, Adrienne", "Lee, Ahlam", 
"Lee, Ahlam; Cloutier, Aimee", "Kleihauer, Jay; Stephens, Roy; Hart, William", 
"Foor, Ryan M.; Cano, Jamie"), pubtitle = c("AIP Conference Proceedings", 
"Journal of Case Studies in Accreditation and Assessment", "173rd Meeting of Acoustical Society of America", 
"Journal of Research in Gender Studies", "Journal of Research in Gender Studies", 
"Scientometrics", "Journal of Agricultural Education", "Journal of Agricultural Education", 
"Journal of Agricultural Education", "Journal of Agricultural Education"
)), class = c("rowwise_df", "tbl_df", "tbl", "data.frame"), row.names = c(NA, 
-10L))

I have a second data frame that has just the author names. Here's a subset of those names, for reproducibility:

structure(list(author = c("Asangba, Abigail", "Stewart, Abigail", 
"Moscatelli, Adriana", "Nishina, Adrienne", "Redmond-Sanogo, Adrienne", 
"Purnamasari, Agustina", "Lee, Ahlam", "Aliyeva, Aida", "Belanger, Aimee", 
"Cloutier, Aimee")), row.names = c(NA, 10L), class = "data.frame")

I'm trying to use this second data frame to subset data from the original data frame, and I'm running into a challenge with the semicolon delimited names.

I thought this would get me there, but no luck so far. I've tried to change the delimited string into a vector and then match against the list of authors, but it only returns names that appear individually (or, I get no matches in names that appear in the string).

list_authors_female <- data %>% 
  select(author, pubtitle) %>% 
  filter(author %in% female_authors_all)

Here, I tried to separate the author column into a vector, but I'm hitting an error.

list_authors_female <- data %>%  
  rowwise() %>% 
  mutate(author_list = str_split(author, pattern = ";")) %>% 
  filter(author_list %in% female_authors_all)

Any pointers? Thanks!

Answer 1

Create a regular expression pat of the form author1|author2|...|authorN and apply it to the pubs . With this approach no splitting is needed.

pat <- authors %>% 
  rowwise %>% 
  mutate(author = toString(author)) %>%
  ungroup %>%
  { paste(.$author, collapse = "|") }

pubs %>% filter(grepl(pat, author))

giving:

# A tibble: 8 x 2
  author                                 pubtitle                               
  <chr>                                  <chr>                                  
1 Moscatelli, Adriana; Nishina, Adrienne AIP Conference Proceedings             
2 Asangba, Abigail                       Journal of Case Studies in Accreditati~
3 Stewart, Abigail                       173rd Meeting of Acoustical Society of~
4 Redmond-Sanogo, Adrienne; Lee, Ahlam   Journal of Research in Gender Studies  
5 Purnamasari, Agustina; Lee, Ahlam; Mo~ Journal of Research in Gender Studies  
6 Nishina, Adrienne                      Scientometrics                         
7 Lee, Ahlam                             Journal of Agricultural Education      
8 Lee, Ahlam; Cloutier, Aimee            Journal of Agricultural Education  

Answer 2

We can use tidyverse approach. Separate the 'author' at the : delimiter into 'long' format, then do an inner_join , later grouped by the row number column already created, paste the 'author' elements back to a single string

library(tidyverse)
df1 %>%
  rownames_to_column('rn') %>% 
  separate_rows(author, sep=";\\s*") %>%
  inner_join(df2)%>% 
  group_by(rn, pubtitle) %>% 
  summarise(author = str_c(author, collapse = "; ")) %>%
  ungroup %>%
  select(names(df1))
# A tibble: 8 x 2
#  author                                                 pubtitle                                               
#  <chr>                                                  <chr>                                                  
#1 Moscatelli, Adriana; Nishina, Adrienne                 AIP Conference Proceedings                             
#2 Asangba, Abigail                                       Journal of Case Studies in Accreditation and Assessment
#3 Stewart, Abigail                                       173rd Meeting of Acoustical Society of America         
#4 Redmond-Sanogo, Adrienne; Lee, Ahlam                   Journal of Research in Gender Studies                  
#5 Purnamasari, Agustina; Lee, Ahlam; Moscatelli, Adriana Journal of Research in Gender Studies                  
#6 Nishina, Adrienne                                      Scientometrics                                         
#7 Lee, Ahlam                                             Journal of Agricultural Education                      
#8 Lee, Ahlam; Cloutier, Aimee                            Journal of Agricultural Education         

---

Or with str_detect and filter

df1 %>% 
    filter(str_detect(author, str_c(df2$author, collapse="|")))

Answer 3

If you're willing to use the tidyr package there are some cool tools for separating out delimited lists. Specifically separate and separate_row .

data
# # A tibble: 10 x 2
#   author                                        pubtitle                                      
#   <chr>                                         <chr>                                         
# 1 Moscatelli, Adriana; Nishina, Adrienne        AIP Conference Proceedings                    
# 2 Asangba, Abigail                              Journal of Case Studies in Accreditation and ~
# 3 Stewart, Abigail                              173rd Meeting of Acoustical Society of America
# 4 Redmond-Sanogo, Adrienne; Lee, Ahlam          Journal of Research in Gender Studies         
# 5 Purnamasari, Agustina; Lee, Ahlam; Moscatell~ Journal of Research in Gender Studies         
# 6 Nishina, Adrienne                             Scientometrics                                
# 7 Lee, Ahlam                                    Journal of Agricultural Education             
# 8 Lee, Ahlam; Cloutier, Aimee                   Journal of Agricultural Education             
# 9 Kleihauer, Jay; Stephens, Roy; Hart, William  Journal of Agricultural Education             
# 10 Foor, Ryan M.; Cano, Jamie                    Journal of Agricultural Education        

female_authors_all
# # A tibble: 10 x 1
#                      author
# 1          Asangba, Abigail
# 2          Stewart, Abigail
# 3       Moscatelli, Adriana
# 4         Nishina, Adrienne
# 5  Redmond-Sanogo, Adrienne
# 6     Purnamasari, Agustina
# 7                Lee, Ahlam
# 8             Aliyeva, Aida
# 9           Belanger, Aimee
# 10          Cloutier, Aimee

data2 <- data %>%
  # If you want to keep the original names duplicate column first
  mutate(author_sep = author) %>%
  # Take each delimited author and give them their own row (tidy data)
  tidyr::separate_rows(author_sep,sep = ";") %>%
  # Filter to only keep rows where the individual author is the other vector
  filter(author_sep %in% female_authors_all$author) %>%
  # Remove that extra column we created
  select(-author_sep) %>%
  # Remove duplicate rows in case more than one author in the delimited list was female
  distinct()

data2
# # A tibble: 8 x 2
#   author                                         pubtitle                                      
#   <chr>                                          <chr>                                         
# 1 Moscatelli, Adriana; Nishina, Adrienne         AIP Conference Proceedings                    
# 2 Asangba, Abigail                               Journal of Case Studies in Accreditation and ~
# 3 Stewart, Abigail                               173rd Meeting of Acoustical Society of America
# 4 Redmond-Sanogo, Adrienne; Lee, Ahlam           Journal of Research in Gender Studies         
# 5 Purnamasari, Agustina; Lee, Ahlam; Moscatelli~ Journal of Research in Gender Studies         
# 6 Nishina, Adrienne                              Scientometrics                                
# 7 Lee, Ahlam                                     Journal of Agricultural Education             
# 8 Lee, Ahlam; Cloutier, Aimee                    Journal of Agricultural Education   

---

Or using inner_join which is more efficient than %in% :

data3 <- data %>%
  # If you want to keep the original names duplicate column first
  mutate(author_sep = author) %>%
  # Take each delimited author and give them their own row (tidy data)
  tidyr::separate_rows(author_sep,sep = ";") %>%
  # inner_join to keep only females
  inner_join(female_authors_all,by = c("author_sep" = "author")) %>%
  # Remove that extra column we created
  select(-author_sep) %>%
  # Remove duplicate rows in case more than one author is the delimited list was female
  distinct()