Scraping a wiki page for the “Periodic table” and all the links

Question

I wish to scrape the following wiki article: http://en.wikipedia.org/wiki/Periodic_table

So that the output of my R code will be a table with the following columns:

Answer 1

Try this:

library(XML)

URL <- "http://en.wikipedia.org/wiki/Periodic_table"
root <- htmlTreeParse(URL, useInternalNodes = TRUE)

# extract attributes and value of all 'a' tags within 3rd table
f <- function(x) c(xmlAttrs(x), xmlValue(x))
m1 <- xpathApply(root, "//table[3]//a", f)
m2 <- suppressWarnings(do.call(rbind, m1))

# extract rows that correspond to chemical symbols
ix <- grep("^[[:upper:]][[:lower:]]{0,2}", m2[, "class"])

m3 <- m2[ix, 1:3]
colnames(m3) <- c("URL", "Name", "Symbol")
m3[,1] <- sub("^", "http://en.wikipedia.org", m3[,1])
m3[,2] <- sub(" .*", "", m3[,2])

A bit of the output:

> dim(m3)
[1] 118   3
> head(m3)
     URL                                      Name        Symbol
[1,] "http://en.wikipedia.org/wiki/Hydrogen"  "Hydrogen"  "H"   
[2,] "http://en.wikipedia.org/wiki/Helium"    "Helium"    "He"  
[3,] "http://en.wikipedia.org/wiki/Lithium"   "Lithium"   "Li"  
[4,] "http://en.wikipedia.org/wiki/Beryllium" "Beryllium" "Be"  
[5,] "http://en.wikipedia.org/wiki/Boron"     "Boron"     "B"   
[6,] "http://en.wikipedia.org/wiki/Carbon"    "Carbon"    "C"   

We can make this more compact by enhancing the xpath expression further starting with Jeffrey's xpath expression (since it nearly gets the elements at top) and adding a qualification to it which exactly does. In that case xpathSApply can be used to eliminate the need for do.call or the plyr package. The last bit where we fix up odds and ends is the same as before. This produces a matrix rather than a data frame which seems preferable since the content is entirely character.

library(XML)

URL <- "http://en.wikipedia.org/wiki/Periodic_table"
root <- htmlTreeParse(URL, useInternalNodes = TRUE)

# extract attributes and value of all a tags within 3rd table
f <- function(x) c(xmlAttrs(x), xmlValue(x))
M <- t(xpathSApply(root, "//table[3]/tr/td/a[.!='']", f))[1:118,]

# nicer column names, fix up URLs, fix up Mercury.
colnames(M) <- c("URL", "Name", "Symbol")
M[,1] <- sub("^", "http://en.wikipedia.org", M[,1])
M[,2] <- sub(" .*", "", M[,2])

View(M)