html_table(..., fill=TRUE) bug in filling rowspans down

[oguzhanogreden]

The bug can be reproduced using an example table from Wikipedia (couldn't find a English example page, this page lists the parliament members of Turkey).

bug <- read_html('https://tr.wikipedia.org/wiki/TBMM_8._d%C3%B6nem_milletvekilleri_listesi') %>%
    html_nodes('.wikitable') %>%
    html_table(fill = TRUE)
bug[[1]][136:142,]

This happens because nrows is not updated when rowspans are filled. For instance in this case, when i == 2 and j == 138, nrows[[j]][i] will be 7, although it should be 1.

This has an easy fix, I'll be creating a PR soon.

[apreshill]

I believe I have just hit on the same issue- here is a reprex from a page written in English:

library(rvest)
#> Loading required package: xml2

url <- "https://en.wikipedia.org/wiki/The_Great_British_Bake_Off"
page <- read_html(url)

series <- page %>%
  html_nodes("table.wikitable") %>%
  .[[1]] %>%
  html_table(fill = TRUE)
#> Error in out[j + k, ]: subscript out of bounds

^{Created on 2019-09-29 by the reprex package (v0.3.0)}

Also looks to be related to this issue on Stack Overflow.