Importing one long line of data into R

Question

I have a large data file consisting of a single line of text. The format resembles

Cat    14  Dog    15  Horse  16

I\'d eventually like to get

Answer 1

Method 1: (extracting from long vector with seq()

> inp <- scan(textConnection("Cat 14 Dog 15 Horse 16"), what="character")
Read 6 items
> data.frame(animal = inp[seq(1,length(inp), by=2)], 
             numbers =as.numeric(inp[seq(2,length(inp), by=2)]))
  animal numbers
1    Cat      14
2    Dog      15
3  Horse      16

Method 2: (using the "what" argument to scan to greater effect)

> inp <- data.frame(scan(textConnection("Cat 14 Dog 15 Horse 16"), 
                     what=list("character", "numeric")))
Read 3 records
> names(inp) <- c("animals", "numbers")
> inp
  animals numbers
1     Cat      14
2     Dog      15
3   Horse      16

This is a refinement of the Method 2: (was worried about possibility of very long column names in the result from scan() so I read the help page again and added names to the what argument values:

inp <- data.frame(scan(textConnection("Cat 14 Dog 15 Horse 16"), 
                        what=list( animals="character", 
                                   numbers="numeric")))
Read 3 records
> inp
  animals numbers
1     Cat      14
2     Dog      15
3   Horse      16

Answer 2

This solution takes full advantage of scan()'s what argument, and seems simpler (to me) than any of the others:

x <- scan(file = textConnection("Cat 14 Dog 15 Horse 16"), 
          what = list(Animal=character(), Number=numeric()))

# Convert x (at this point a list) into a data.frame
as.data.frame(x)
#   Animal Number
# 1    Cat     14
# 2    Dog     15
# 3  Horse     16

Answer 3

Assuming that the white space is a delimiter, you can use the following mechanism:

• Use scan to read the file
• Convert the results to a matrix, then to a data.frame

The code:

x <- scan(file=textConnection("
Cat 14 Dog 15 Horse 16
"), what="character")

xx <- as.data.frame(matrix(x, ncol=2, byrow=TRUE))
names(xx) <- c("Animal", "Number")
xx$Number <- as.numeric(xx$Number)

The results:

xx

  Animal Number
1    Cat      1
2    Dog      2
3  Horse      3

Answer 4

Here is another approach

string <- readLines(textConnection(x))
string <- gsub("(\\d+)", "\\1\n", string, perl = TRUE)
dat    <- read.table(text = string, sep = "")

Answer 5

One way:

# read the line
r <- read.csv("exa.Rda",sep=" ", head=F)
# every odd number index is an animal
animals <- r[,(1:ncol(r)-1)%%2==0]
# every even number index is a number
numbers <- r[,(1:ncol(r))%%2==0]
# flipping the animal row into a column
animals <- t(animals)
# flipping the number row into a column
numbers <- t(numbers)
# putting the data together
mydata <- data.frame(animals, numbers)

Answer 6

Here's one solution using a variety of tools/hacks, specifically:

• strplit to split on space characters (\\s)
• unlist to coerce the list returned by strsplit into a vector
• matrix to turn the vector into the appropriate shape
• data.frame to allow for columns of different mode
• as.character and as.numeric to convert the Count column from a factor

Here's everything put together:

txt <- "Cat 14 Dog 15 Horse 16"

out <- data.frame(matrix(unlist(strsplit(txt, "\\s")), ncol = 2, byrow = TRUE, dimnames = list(NULL, c("Animal", "Count"))))
out$Count <- as.numeric(as.character(out$Count))
str(out)

'data.frame':   3 obs. of  2 variables:
 $ Animal: Factor w/ 3 levels "Cat","Dog","Horse": 1 2 3
 $ Count : num  14 15 16