R converting XML nodesets of <p> elements into dataframe when scraping table with rvest-CodePudding

We are pulling contact info from https://nbpa.com/agents/directory. There is no table on the page, but rather <div>s with <p> elements inside:

We can grab this element with:

agents_url <- "https://nbpa.com/agents/directory"
agents_page <- agents_url %>% read_html()
agents_page_elements <- agents_page %>% html_nodes('div.accordion-inner')

agents_page_elements[1]
agents_page_elements[1] %>% html_nodes('p')

We are looking to convert this into a 1-row dataframe:

Cell             Email                  Professional Credentials:
(123) 456-7890   firstlast@email.com    "NBA Certified Player Agent..."

Is this possible to do? The challenging part about this web scrape is that each accordion-inner div on the website has different p elements. Some have Cell: and Email:, others have Education:, Address:, etc. It varies by accordion-inner. If we can turn each individual node into a 1-row dataframe, we can then rbind all dataframes together using plyr::rbind.fill().

CodePudding user response：

We can use read.dcf after getting as text

new <- agents_page_elements[1] %>%
     html_nodes('p') %>%
     html_text()
as.data.frame(read.dcf(textConnection(new)))

-output

    Cell                            Email                                                             Professional Credentials
1 (240) 668-4241 barry.aberdeen@tributesports.com NBA Certified Player Agent, FIBA Certified Player Agent, WNBA Certified Player Agent

For multiple elements, use map

library(purrr)
library(dplyr)
library(stringr)
out <- map_dfr(agents_page_elements, ~ {
       new <- .x %>% 
               html_nodes('p') %>%
               html_text() %>%
               str_replace_all("\n\\s*", " ")
    if(length(new) > 0) {
       as.data.frame(read.dcf(textConnection(new)))
     } else NULL

     })

-output

> dim(out)
[1] 455   9
> head(out, 2)
            Cell                            Email                                                             Professional Credentials    Title
1 (240) 668-4241 barry.aberdeen@tributesports.com NBA Certified Player Agent, FIBA Certified Player Agent, WNBA Certified Player Agent     <NA>
2 (281) 773-7339                             <NA>                                     Texas Bar No. 24050197|Wisconsin Bar No. 1045470 Attorney
             Company Name                                                                                             Education Address Office International
1                    <NA>                                                                                                  <NA>    <NA>   <NA>          <NA>
2 Adams & Associates, LLC University of Southern California B.S. |University of Houston - M.B.A |University of Wisconsin - J.D.    <NA>   <NA>          <NA>