How to split dataframe column in R

Question

I have a column in a dataframe called parm_value that I would like to split into two columns, lower and upper bound based on the position of the underscore in the field. I have been trying to use a combination of grep and substring with no success

Current dataframe format:

   parm_value
1       30_34
2       60_64
3       65_69
4       75_79
5       90_94

Desired data frame format:

   parm_value   lower_bound   upper_bound
1       30_34            30            34
2       60_64            60            64
3       65_69            65            69
4       75_79            75            79
5       90_94            90            94

I have been trying things like

dat02 <-
   dat01 %>%
   mutate(lower_bound = substring(parm_value, 1, grep("_", parm_value) - 1)

Answer 1

You can try if your data.frame is called df :

cbind(df, `colnames<-`( do.call("rbind", sapply(df[,1], strsplit, "_")), c("lower bound", "upper bound")))

  #    parm_value lower bound upper bound
  #  1      30_34          30          34
  #  2      60_64          60          64
  #  3      65_69          65          69
  #  4      75_79          75          79
  #  5      90_94          90          94

Answer 2

Use strsplit :

library(data.table)
xmpl <- data.table(val = rep("65_45", 5))
xmpl[ , lower := sapply(strsplit(val, "_"), "[[", 1)]
xmpl[ , upper := sapply(strsplit(val, "_"), "[[", 2)]

xmpl
#      val lower upper
# 1: 65_45    65    45
# 2: 65_45    65    45
# 3: 65_45    65    45
# 4: 65_45    65    45
# 5: 65_45    65    45

If it is a really large table you can save runtime by only running the strsplit once, then calling the object when defining the new data.table fields.

strsplit returns a list:

strsplit("65_45", "_")
# [[1]]
# [1] "65" "45"

The sapply call iterates through the list with the subsetting function [[ selecting the Nth item, where N is given in sapply as sapply(some_list, "[[", N) .

Answer 3

you could also use cSplit from splitstackshape

library(splitstackshape)
out = cbind(dat, setnames(cSplit(dat, "parm_value", "_", fixed = FALSE),
      c("lower_bound", "upper_bound")))

#> out
#  parm_value lower_bound upper_bound
#1      30_34          30          34
#2      60_64          60          64
#3      65_69          65          69
#4      75_79          75          79
#5      90_94          90          94

Answer 4

Try read.table

cbind(df1[1],read.table(text= as.character(df1$parm_value), sep="_", 
             col.names=c('lower_bound', 'upper_bound')))
#    parm_value lower_bound upper_bound
#1      30_34          30          34
#2      60_64          60          64
#3      65_69          65          69
#4      75_79          75          79
#5      90_94          90          94

Or separate from tidyr

library(tidyr)
separate(df1, parm_value, into=c('lower_bound', 'upper_bound'), remove=FALSE)
#    parm_value lower_bound upper_bound
#1      30_34          30          34
#2      60_64          60          64
#3      65_69          65          69
#4      75_79          75          79
#5      90_94          90          94