rvest: select and scrape html table(s) after a specific (title) string - r

Question

I am trying to scrape the following website with public healthcare data:
https://prog.nfz.gov.pl/app-jgp/Grupa.aspx?id=Qpc6nYOpOBQ%3d
I would like to scrape only a table or tables that are preceded with a title
"Tabela xx procedury ICD-9" where xx is not a fixed number.
There may be 1 but occasionally 2-3 such tables on the page. And they can occur in different order with other pages, so it is not possible to indicate I want n'th page on the website.
For instance, I am interested only in tables in the same page that are preceded by "Icd-9 main" string and skip others. There are 2 such tables and I would like to scrape their content to a data.frame. They may not appear in the same order as below, so I must rely on preceding string. Occasionally there will be no Icd-9 main table at all.
Page
----
Icd-10
====
Table
====
Icd-9 main
====
Table
===
Icd-9 main
====
Table
====
Icd-9 supplementary
====
Table
===
I am only aware of the code that selects the nth table in order like in this tutorial:
https://www.r-bloggers.com/using-rvest-to-scrape-an-html-table/
library("rvest")
url <- "http://en.wikipedia.org/wiki/List_of_U.S._states_and_territories_by_population"
population <- url %>%
html() %>%
html_nodes(xpath='//*[#id="mw-content-text"]/table[1]') %>%
html_table()
population <- population[[1]]
whereby all the tables are dumped into a list and then we can select which table we want by number
However in my case I never know which table to scrape, the order may vary, it can be 2 tables preceded by a string containing "Tabela xx: procedury ICD-9"
My question is, how to select and scrape html tables based on preciding title or description string. Or a table or tables appearing only immediately after a string containing e.g "Tabela xx procedury ICD-9"

Answer 1

library(rvest)
library(stringr)
doc <- read_html("https://prog.nfz.gov.pl/app-jgp/Grupa.aspx?id=Qpc6nYOpOBQ%3d")
# extract all the nodes that have the title (id = "tytul") or a table
# the cs selector "," is like a boolean OR.
nodes <- doc %>% html_nodes(".tytul,table")
# loop though each node.
signal <- FALSE
my_tables <- list()
j <- 0
for (i in 1:length(nodes)) {
# if title signal previously set and this is a table tag
if (signal & html_name(nodes[i]) == "table") {
cat("Match..\n")
# get the table (data frame)
this_table <- html_table(nodes[i], fill = TRUE, header = TRUE)[[1]]
# append to list
j = j + 1
my_tables[[j]] <- this_table
# and reset the signal so we search for the next one
signal <- FALSE
}
# if the signal is clear look for matching title
if (!signal) {
signal <- nodes[i] %>% html_text() %>% str_detect("Tabela.+ICD 9")
}
}
my_tables[[1]][1:5,]
my_tables[[2]][1:5,]
# > my_tables[[1]][1:5,]
# ICD 9 Nazwa Lb. hospitalizacji UdziaĹ\u0082 (%) Mediana czasu pobytu (dni)
# 1 2.051 ZaĹ\u0082oĹźenie pĹ\u0082ytki sztucznej do czaszki 168 32,31 7
# 2 1.247 Kraniotomia z usuniÄ\u0099ciem krwiaka podtwardĂłwkowego 55 10,58 20
# 3 2.022 Odbarczenie zĹ\u0082amania czaszki 43 8,27 6
# 4 2.040 Przeszczep kostny do koĹ\u009bci czaszki 35 6,73 8
# 5 1.093 Inne aspiracje w zakresie czaszki 33 6,35 5
# > my_tables[[2]][1:5,]
# ICD 9 Nazwa Lb. hospitalizacji UdziaĹ\u0082 (%) Mediana czasu pobytu (dni)
# 1 O35 SĂłd (Na) 239 45,96 8
# 2 89.00 Porada lekarska, konsultacja, asysta 230 44,23 9
# 3 N45 Potas (K) 217 41,73 8
# 4 87.030 TK gĹ\u0082owy bez kontrastu 214 41,15 9
# 5 89.04 Opieka pielÄ\u0099gniarki lub poĹ\u0082oĹźnej 202 38,85 8