在 R 和 rvest 中抓取多个链接的 HTML 表

Question

这篇文章http://www.ajnr.org/content/30/7/1402.full 包含四个指向我想用rvest 抓取的html 表的链接。

借助 css 选择器：

"#T1 a" 

可以像这样到达第一张桌子：

library("rvest")
html_session("http://www.ajnr.org/content/30/7/1402.full") %>%
follow_link(css="#T1 a") %>%
html_table() %>%
View()

css 选择器：

".table-inline li:nth-child(1) a"

可以选择包含链接到四个表的标签的所有四个 html 节点：

library("rvest")
html("http://www.ajnr.org/content/30/7/1402.full") %>%
html_nodes(css=".table-inline li:nth-child(1) a")

如何遍历此列表并一次检索所有四个表？最好的方法是什么？

Answer 1

你可能想使用如下：

main_url <- "http://www.ajnr.org/content/30/7/1402/"
urls <- paste(main_url,c("T1.expansion","T2.expansion","T3.expansion","T4.expansion"),".html", sep = "")
tables <- list()
for(i in seq_along(urls))
{
  total <- readHTMLTable(urls[i])
  n.rows <- unlist(lapply(total, function(t) dim(t)[1]))
  tables[[i]] <- as.data.frame(total[[which.max(n.rows)]])
}
tables

#[[1]]
#  Glioma Grade Sensitivity Specificity    PPV    NPV
#1    II vs III       50.0%       92.9%  80.0%  76.5%
#2     II vs IV      100.0%      100.0% 100.0% 100.0%
#3    III vs IV       78.9%       87.5%  93.8%  63.6%

#[[2]]
#  Glioma Grade Sensitivity Specificity   PPV    NPV
#1    II vs III       87.5%       71.4% 63.6%  90.9%
#2     II vs IV      100.0%       85.7% 90.5% 100.0%
#3    III vs IV       89.5%       75.0% 89.5%  75.0%

#[[3]]
#  Criterion Sensitivity Specificity    PPV   NPV
#1       ≥1*       85.2%       92.9%  95.8% 76.5%
#2        ≥2       81.5%      100.0% 100.0% 73.7%

#[[4]]
#  Criterion Sensitivity Specificity   PPV   NPV
#1     <1.92       96.3%       71.4% 86.7% 90.9%
#2     <2.02       92.6%       71.4% 86.2% 83.3%
#3    <2.12*       92.6%       85.7% 92.6% 85.7%

Answer 2

这是一种方法：

library(rvest)

url <- "http://www.ajnr.org/content/30/7/1402.full"
page <- read_html(url)

# First find all the urls
table_urls <- page %>% 
  html_nodes(".table-inline li:nth-child(1) a") %>%
  html_attr("href") %>%
  xml2::url_absolute(url)

# Then loop over the urls, downloading & extracting the table
lapply(table_urls, . %>% read_html() %>% html_table())