Split Column By Space

February 28, 2022

DATA <- data.frame("V1" = c("SCORE : 9.931 5.092",
                            "SCORE : 6.00007 15.1248",
                            "SCORE : 1.0002 12.987532",
                            "SCORE : 3.1 3.98532"))
WANT <- data.frame("VAR1" = c(9.931, 6.00007, 1.0002, 3.1),
                   'VAR2' = c(5.092, 15.1248, 12.987532, 3.98532))

What I have is student test score data that was entered like shown in ‘DATA’ but I wish to split it so that what I have is like what is shown in the ‘WANT’ frame where the first number is in ‘VAR1’ and the second number is in ‘VAR2’ ignoring the spaces

MY ATTEMPT:

DATA[, c("VAR1", "VAR2") := trimws(V1, whitespace = ".*:\\s+")]

PRODUCE:

>Solution :

We can remove the prefix substring with trimws and read the column using read.table with sep as default space to return two column data.frame in base R

read.table(text = trimws(DATA$V1, whitespace = ".*:\\s+"), 
   header = FALSE, col.names = c("VAR1", "VAR2"))
     VAR1     VAR2
1 9.93100  5.09200
2 6.00007 15.12480
3 1.00020 12.98753
4 3.10000  3.98532

Or may use extract from tidyr

library(tidyr)
extract(DATA, V1, into = c("VAR1", "VAR2"),
    ".*:\\s+([0-9.]+)\\s+([0-9.]+)", convert = TRUE)
     VAR1     VAR2
1 9.93100  5.09200
2 6.00007 15.12480
3 1.00020 12.98753
4 3.10000  3.98532

If we want data.table, using the same method can read with fread after removing the prefix substring

library(data.table)
fread(text = setDT(DATA)[, trimws(V1, whitespace = ".*:\\s+")], 
   col.names = c("VAR1", "VAR2"))
      VAR1     VAR2
     <num>    <num>
1: 9.93100  5.09200
2: 6.00007 15.12480
3: 1.00020 12.98753
4: 3.10000  3.98532

Or use the select option in fread

fread(text = DATA$V1, select = c(3, 4), col.names = c("VAR1", "VAR2"))
     VAR1     VAR2
     <num>    <num>
1: 9.93100  5.09200
2: 6.00007 15.12480
3: 1.00020 12.98753
4: 3.10000  3.98532

Or read as four columns, and subset

fread(text = DATA$V1)[, .(VAR1 = V3, VAR2 = V4)]
     VAR1     VAR2
     <num>    <num>
1: 9.93100  5.09200
2: 6.00007 15.12480
3: 1.00020 12.98753
4: 3.10000  3.98532

Or can use tstrsplit

setDT(DATA)[, c("VAR1", "VAR2") := tstrsplit(trimws(V1, 
       whitespace = ".*:\\s+"), " ")]
DATA <- type.convert(DATA, as.is = TRUE)
DATA
                         V1    VAR1     VAR2
                     <char>   <num>    <num>
1:      SCORE : 9.931 5.092 9.93100  5.09200
2:  SCORE : 6.00007 15.1248 6.00007 15.12480
3: SCORE : 1.0002 12.987532 1.00020 12.98753
4:      SCORE : 3.1 3.98532 3.10000  3.98532