CODEWITHSUNDEEP
rdata.tablestringi
Crops
Crops

Reputation: 5162

stri_split_fixed in a data.table in R

I have a data.table DT as follows.

DT <- structure(list(V1 = structure(1:3, .Label = c("S01", "S02", "S03" ), class = "factor"), V2 = structure(c(1L, 3L, 2L), .Label = c("Alan Hal << Guy John", "Bruce Dick Jean-Paul << Damien", "Jay << Barry Wally Bart"), class = "factor")), .Names = c("V1", "V2"), row.names = c(NA, -3L), class = "data.frame")
# DT
#    V1                             V2
# 1 S01           Alan Hal << Guy John
# 2 S02        Jay << Barry Wally Bart
# 3 S03 Bruce Dick Jean-Paul << Damien
setDT(DT)

I am trying to split the column V2 at "<<" and the get the output in two new columns.

I could get it done as follows using stringi

T <- as.data.frame(do.call(rbind,  stri_split_fixed(DT$V2, "<<", 2)))
setnames(T, old = colnames(T), new = c("V3", "V4"))
cbind(DT, T)
V1                             V2                    V3                V4
1: S01           Alan Hal << Guy John             Alan Hal           Guy John
2: S02        Jay << Barry Wally Bart                  Jay   Barry Wally Bart
3: S03 Bruce Dick Jean-Paul << Damien Bruce Dick Jean-Paul             Damien

However I would like to do the same by reference using the := operator. How to do this using data.table?

I am having difficulty with the RHS part. 

DT[, c("V1", "V2) := list()]

stri_split_fixed(DT$V2, "<<", 2) gives a list of 3 with character vectors of length 2. How to get a list of 2 with character vectors of length 3?

Upvotes: 1

Views: 676

Answers (1)

akrun
akrun

Reputation: 886948

You could try

setDT(DT)[, c('V3', 'V4'):=do.call(rbind.data.frame,
                    stri_split_fixed(V2, ' << ', 2))][]
#  V1                             V2                    V3                V4
#1: S01           Alan Hal << Guy John             Alan Hal           Guy John
#2: S02        Jay << Barry Wally Bart                  Jay   Barry Wally Bart
#3: S03 Bruce Dick Jean-Paul << Damien Bruce Dick Jean-Paul             Damien

Or you could use strsplit (from @David Arenburg's comments)

 setDT(DT)[, c('V3', 'V4'):= do.call(rbind.data.frame,
                   strsplit(as.character(V2), " << "))]

More efficient option (as suggested by @Ananda Mahto)

cbind(DT, `colnames<-`(stri_split_fixed(DT$V2,
              " << ", simplify = TRUE), c("V3", "V4")))

Another option would be to use cSplit from splitstackshape

library(splitstackshape)
cSplit(DT, 'V2', ' << ', stripWhite=FALSE, drop=FALSE)
#       V1                             V2                 V2_1             V2_2
#1: S01           Alan Hal << Guy John             Alan Hal         Guy John
#2: S02        Jay << Barry Wally Bart                  Jay Barry Wally Bart
#3: S03 Bruce Dick Jean-Paul << Damien Bruce Dick Jean-Paul           Damien

A faster version of cSplit which gives similar performance as stri_split is available in Gist

Upvotes: 6

Related Questions