1

I am working on scraping some data into a data frame, and am getting some empty fields, where I would instead prefer to have NA. I have tried na.strings, but am either placing it in the wrong place or it just isn't working, and I tried to gsub anything that was whitespace from beginning of line to end, but that didn't work. 

htmlpage <- read_html("http://www.gourmetsleuth.com/features/wine-cheese-pairing-guide")
sugPairings <- html_nodes(htmlpage, ".meta-wrapper")
suggestions <- html_text(sugPairings)
suggestions <- gsub("\\r\\n", '', suggestions)

How can I sub out the blank fields with NA, either once it is added to the data frame, or before adding it.

Jamie Leigh
  • 359
  • 1
  • 4
  • 18

1 Answers1

2

rvest::html_text has an build in trimming option setting trim=TRUE. After you have done this you can use e.g. ifelse to test for an empty string (=="") or use nzchar.

I full you could do this:

html_nodes(htmlpage, ".meta-wrapper") %>% html_text(trim=TRUE) %>% ifelse(. == "", NA, .)

or this:

res <- html_nodes(htmlpage, ".meta-wrapper") %>% html_text(trim=TRUE)
res[!nzchar(res)] <- NA_character_

@Richard Scriven improvement:

html_nodes(htmlpage, ".meta-wrapper") %>% html_text(trim=TRUE) %>% replace(!nzchar(.), NA)
Rentrop
  • 20,979
  • 10
  • 72
  • 100