this is the code i wrote to generate the data:

info <- html_nodes(manga, ".mt4") %>% html_text2() %>% strsplit("\n")

it returns 50 rows of lists that that look like this:

[1] "Manga (? vols)" "Aug 1989 -" "741,705 members"

without the strsplit function it returns this kind of string:

[1] "Manga (? vols)\nAug 1989 -\n741,705 members"

my end goal is to have 3 separate columns, like below but with 50 rows:

this is the code i wrote to generate the data:

info <- html_nodes(manga, ".mt4") %>% html_text2() %>% strsplit("\n")

it returns 50 rows of lists that that look like this:

[1] "Manga (? vols)" "Aug 1989 -" "741,705 members"

without the strsplit function it returns this kind of string:

[1] "Manga (? vols)\nAug 1989 -\n741,705 members"

my end goal is to have 3 separate columns, like below but with 50 rows:

a complete table would look like this, i already have the title and rating section figured out:

im not sure if i should approach the question by keeping my data as rows of strings and trying to split it into columns that way, or figure out what i can do to a rows of lists. i have tried to see if the css selector can be more specific, but it seems like MyAnimeList just lumps all that information together so im not sure what to do

• r
• web-scraping
• rvest

• what is the definition of manga? – s_baldur Commented Mar 23 at 9:50

3 Answers 3

I assume it's https://myanimelist/topmanga.php you are working on.

This is how text is anized in div.mt4:

mt4 <- rvest::html_element(rvest::read_html("https://myanimelist/topmanga.php"), ".mt4")
as.character(mt4) %>% cat()
#> <div class="information di-ib mt4">
#>         Manga (? vols)<br>
#>         Aug 1989 - <br>
#>         741,735 members
#>       </div>
xml2::html_structure(mt4)
#> <div.information.di-ib.mt4>
#>   {text}
#>   <br>
#>   {text}
#>   <br>
#>   {text}

So you can access each individual text section separated by <br> with xpath, if you choose to:

html_element(mt4, xpath = "./text()[1]") %>%  html_text2()
#> [1] "Manga (? vols)"
html_element(mt4, xpath = "./text()[2]") %>% html_text2()
#> [1] "Aug 1989 -"
html_element(mt4, xpath = "./text()[3]") %>% html_text2()
#> [1] "741,735 members"

Though I personally would not bother unless there are known inconsistencies, just collect and separate at newlines:

library(rvest)
library(tidyr)

read_html("https://myanimelist/topmanga.php") %>%
  html_elements("tr.ranking-list") %>% 
  {
    tibble(
      title  = html_element(., "h3.manga_h3") %>% html_text(),
      rating = html_element(., ".score span") %>% html_text(),
      info   = html_element(., ".mt4") %>% html_text2()
    )
  } %>%
  separate_wider_delim(info, "\n", names = c("media", "serialization", "members"))
#> # A tibble: 50 × 5
#>    title                                      rating media serialization members
#>    <chr>                                      <chr>  <chr> <chr>         <chr>  
#>  1 Berserk                                    9.47   Mang… Aug 1989 -    741,73…
#>  2 JoJo no Kimyou na Bouken Part 7: Steel Ba… 9.32   Mang… Jan 2004 - A… 287,98…
#>  3 Vagabond                                   9.26   Mang… Sep 1998 - M… 417,71…
#>  4 One Piece                                  9.22   Mang… Jul 1997 -    654,63…
#>  5 Monster                                    9.16   Mang… Dec 1994 - D… 264,90…
#>  6 Slam Dunk                                  9.08   Mang… Sep 1990 - J… 183,84…
#>  7 Vinland Saga                               9.08   Mang… Apr 2005 -    327,82…
#>  8 Fullmetal Alchemist                        9.04   Mang… Jul 2001 - S… 305,86…
#>  9 Tian Guan Cifu                             9.03   Nove… Feb 2021 - M… 14,678…
#> 10 Grand Blue                                 9.03   Mang… Apr 2014 -    185,87…
#> # ℹ 40 more rows

^{Created on 2025-03-23 with reprex v2.1.1}

Like others I also assume you are referring to /topmanga.php. Here is one way to convert your list of string vectors into a dataframe in base R.

The approach below uses lapply to iterate over the t() call (transpose), passing the output of the transposition into as.data.frame(). This returns a list of dataframes. Then, the list of dataframes is bound together using rbind within do.call().

library(rvest)

url <- "https://myanimelist/topmanga.php"

manga <- read_html(url)

df <-
    html_nodes(manga, ".mt4") %>%
    html_text2() %>%
    strsplit("\n") %>%
    lapply(\(listElement) as.data.frame(t(listElement))) %>%
    do.call(rbind, .)

head(df)

Printing out:

               V1                  V2              V3
1  Manga (? vols)          Aug 1989 - 741,744 members
2 Manga (24 vols) Jan 2004 - Apr 2011 287,986 members
3 Manga (37 vols) Sep 1998 - May 2015 417,716 members
4  Manga (? vols)          Jul 1997 - 654,636 members
5 Manga (18 vols) Dec 1994 - Dec 2001 264,904 members
6 Manga (31 vols) Sep 1990 - Jun 1996 183,847 members

Note that if you wish to use options other than base R, then dplyr::bind_rows() can be used instead of do.call(rbind, .).

Having shared all that for how to deal with your strsplit() output, I find the answer by margusl to probably be more principled for your broader objective.

Thanks to @Margusl for providing the link. In this case it can also be viable to read the html table directly using readHTMLTable(url) and unnest_wider the Title-column that includes multiple information by \n.

library(XML)
library(RCurl)
library(rlist)
library(tidyverse)
url <- getURL("https://myanimelist/topmanga.php",.opts = list(ssl.verifypeer = FALSE) )
tables <- readHTMLTable(url) # gives all tables on the page

df <- tables[[1]] %>%
  mutate(Title = str_split(Title, "\\n+")) %>% # split Title column  by `\n`
  unnest_wider(Title, names_sep = "_") %>%
  mutate(across(where(is.character), ~ trimws(.x))) # trim

kableExtra::kable(head(df))

giving