Split column into multiple columns R

Question

I have a data frame column that I need to split into 3 separate column. Looks like this:

I:500-600
I:700-900
II:200-250

I'd like to split this into the following 3 columns:

V1 V2 V3
I 500 600
I 700 900
II 200 250

This has proved slightly trickier than I had hoped. Any help would be appreciated.

Solution 1

Another solution with str_match from the stringr package:

x <- c("I:500-600", "I:700-900", "II:200-250")
library(stringr)
as.data.frame(str_match(x, "^(.*):(.*)-(.*)$")[,-1])
##   V1  V2  V3
## 1  I 500 600
## 2  I 700 900
## 3 II 200 250

In the above regular expression we match 3 substrings: from the beginning to :, from : to -, and from - to the end. Each matched substring will constitute a separate column in the resulting object.

Solution 2

You can use strsplit with an OR argument splitting using : or - this will give you a list which you can process further.

> test <- c('I:500-600', 'I:700-900', 'II:200-250')
> do.call(rbind.data.frame, strsplit(test, ":|-"))
  c..I....I....II.. c..500....700....200.. c..600....900....250..
1                 I                    500                    600
2                 I                    700                    900
3                II                    200                    250

If names are important

> as.data.frame(do.call(rbind, strsplit(test, ":|-")))
  V1  V2  V3
1  I 500 600
2  I 700 900
3 II 200 250

Solution 3

Other options include extract from tidyr

library(tidyr)
extract(df1, V1, into=c('V1','V2', 'V3'),
            '([^:]*):([0-9]*)-([0-9]*)', convert=TRUE)
#  V1  V2  V3
#1  I 500 600
#2  I 700 900
#3 II 200 250

Or tstrsplit from data.table.

library(data.table)#v1.9.5+
setDT(df1)[, tstrsplit(V1, '[:-]', type.convert=TRUE)]
#   V1  V2  V3
#1:  I 500 600
#2:  I 700 900
#3: II 200 250

NOTE: Both options have arguments to convert the class of the output columns

data

df1 <- structure(list(V1 = c("I:500-600", "I:700-900", "II:200-250")), 
 .Names = "V1", class = "data.frame", row.names = c(NA, -3L))

Solution 4

I would recommend cSplit from my "splitstackshape" package.

The syntax is pretty straightforward: cSplit(yourInputDataFrame, yourSplittingColumn, theDelimiters).

Here's an example on a vector. You'd skip the data.table part if you already had a data.frame or a data.table.

library(splitstackshape)
cSplit(data.table(x), "x", ":|-", fixed = FALSE)
#    x_1 x_2 x_3
# 1:   I 500 600
# 2:   I 700 900
# 3:  II 200 250

By default, it also runs type.convert:

str(.Last.value)
# Classes ‘data.table’ and 'data.frame':    3 obs. of  3 variables:
#  $ x_1: Factor w/ 2 levels "I","II": 1 1 2
#  $ x_2: int  500 700 200
#  $ x_3: int  600 900 250
#  - attr(*, ".internal.selfref")=<externalptr> 

View more solutions