Split text string in a data.table columns

Update: From version 1.9.6 (on CRAN as of Sep'15), we can use the function tstrsplit() to get the results directly (and in a much more efficient manner):

require(data.table) ## v1.9.6+
dt[, c("PX", "PY") := tstrsplit(PREFIX, "_", fixed=TRUE)]
#    PREFIX VALUE PX PY
# 1:    A_B     1  A  B
# 2:    A_C     2  A  C
# 3:    A_D     3  A  D
# 4:    B_A     4  B  A
# 5:    B_C     5  B  C
# 6:    B_D     6  B  D

tstrsplit() basically is a wrapper for transpose(strsplit()), where transpose() function, also recently implemented, transposes a list. Please see ?tstrsplit() and ?transpose() for examples.

See history for old answers.

Using splitstackshape package:

library(splitstackshape)
cSplit(df, splitCols = "PREFIX", sep = "_", direction = "wide", drop = FALSE)
#    PREFIX VALUE PREFIX_1 PREFIX_2
# 1:    A_B     1        A        B
# 2:    A_C     2        A        C
# 3:    A_D     3        A        D
# 4:    B_A     4        B        A
# 5:    B_C     5        B        C
# 6:    B_D     6        B        D

We could try:

library(data.table)  
cbind(dt, fread(text = dt$PREFIX, sep = "_", header = FALSE))
    #    PREFIX VALUE V1 V2
    # 1:    A_B     1  A  B
    # 2:    A_C     2  A  C
    # 3:    A_D     3  A  D
    # 4:    B_A     4  B  A
    # 5:    B_C     5  B  C
    # 6:    B_D     6  B  D

I add answer for someone who do not use data.table v1.9.5 and also want an one line solution.

dt[, c('PX','PY') := do.call(Map, c(f = c, strsplit(PREFIX, '-'))) ]