The R Book

The readLnes Function

In some cases you might want to read each line from a file separately. The argument n=-1 means read to the end of the file. Let's try it out with the murders data (p. 100):

readLines("c:\\temp\\murders.txt",n=-1)

This produces the rather curious object of class = "character":

[1] "state\tpopulation\tmurder\tregion"  "Alabama\t3615\t15.1\tSouth"
[3] "Alaska\t365\t11.3\tWest"            "Arizona\t2212\t7.8\tWest"

.......
........
[49] "West.Virginia\t1799\t6.7\tSouth"   "Wisconsin\t4589\t3\tNorth.Central"
[51] "Wyoming\t376\t6.9\tWest"

Each line has been converted into a single character string. Line [1] contains the four variable names (see above) separated by tab characters \t. Line [2] contains the first row of data for murders in Alabama, while row [51] contains the last row of data for murders in Wyoming. You can use the string-splitting function strsplit to tease apart the elements of the string (say, for the Wyoming data [51]):

mo<-readLines("c:\\temp\\murders.txt",n=-1)
strsplit(mo[51],"\t")

[[1]]
[1] "Wyoming""376"  "6.9" "West"

You would probably want 376 and 6.9 as numeric rather than character objects:

as.numeric(unlist(strsplit(mo[51],"\t")))

[1]  NA  376.0 6.9  NA
Warning message:
NAs  introduced by coercion

where the two names Wyoming and West have been coerced to NA, or

as.vector(na.omit(as.numeric(unlist(strsplit(mo[51],"\t")))))

[1] 376.0  6.9
Warning message:
NAs introduced by coercion

to get the numbers on their own. Here is how to extract the two ...