CODEWITHSUNDEEP
r
mendy
mendy

Reputation: 359

Match substrings to strings in a list and record each matched index

I have a dataframe of substrings and a list of strings. I want to check which substrings match which element and record the list indices of matches in the dataframe.

my_list <- list("hello there", "how are you?", "I am fine thanks")

words <- data.frame(text = c("he", "she", "they", "you", "I"), index = NA)

The final output should be:

> words
  text index
1   he    NA
2  she    NA
3 they    NA
4  you     2
5    I     3

I've tried a loop with grepl, which failed to work both by recording the contents instead of the index, and by not recording the correct element:

for (i in 1:nrow(words)){
  x <- grepl(words$text[i], my_list, fixed = T)
  if (x == T) {
    words$index[i] <- paste(my_list[i])
  }
}

> words
  text       index
1   he hello there
2  she        <NA>
3 they        <NA>
4  you        <NA>
5    I        <NA>

I also tried this answer which looked good but which only returned a vector of FALSEs as long as my_list.

EDIT: I'm a bit closer now with this loop, although it's still indexing "he" incorrectly due to the "_he_llo there".

for (i in seq_along(my_list)){
  for (j in 1:nrow(words)){
    if (grepl(words$text[j], my_list[i], fixed = T) == T){
      words$index[[j]] <- i
    }
  }
}
> words
  text index
1   he     1
2  she    NA
3 they    NA
4  you     2
5    I     3

So, how can I match the element? And then, how can I record the matched element's index?

Thanks!

Upvotes: 1

Views: 82

Answers (1)

tmfmnk
tmfmnk

Reputation: 40131

One solution involving dplyr, tidyr, string and purrr could be:

map2_dfr(.x = my_list,
         .y = 1:length(my_list),
         ~ set_names(str_detect(.x, paste0("\\b", words$text, "\\b")) * .y, words$text)) %>%
 summarise_all(max) %>%
 pivot_longer(everything(), names_to = "text", values_to = "index")

  text  index
  <chr> <int>
1 he        0
2 she       0
3 they      0
4 you       2
5 I         3

Or if you want NAs:

map2_dfr(.x = my_list,
         .y = 1:length(my_list),
         ~ set_names(str_detect(.x, paste0("\\b", words$text, "\\b")) * .y, words$text)) %>%
 summarise_all(~ if (all(. == 0)) NA else max(.)) %>%
 pivot_longer(everything(), names_to = "text", values_to = "index")

  text  index
  <chr> <int>
1 he       NA
2 she      NA
3 they     NA
4 you       2
5 I         3

Upvotes: 1

Related Questions