Modify value of column in data.table without loop

Question

I have a datatable which looks like the following

> head(mydt)
    name  b      c
1:  ao    2      1 GiB
2:  bo    2      1.4 Gib

Now, I try to do some cleansing - i try to remove the unit from the values in column c without any loops,.. I did the following:

mydt[,4 :=substr(c,0,gregexpr(pattern=' ',c)[[1]][1]-1)]

What I get is something like this:

> head(mydt)
    name  b      c
1:  ao    2      1 G
2:  bo    2      1.4

What I expect is the following

> head(mydt)
    name  b      c
1:  ao    2      1
2:  bo    2      1.4

However, it doesn't work --> it seems to use the same endpoint for all values. What I am doing wrong? How can I access the "current" value?

my question is not about extracting numbers from vectors -> it is about how to access the current value in order to modify it! — user3579222
– user3579222, Commented Dec 17, 2018 at 21:08
"i try to remove the unit from the values in column c" can easily be interpreted as that you want to remove the unit, i.e. extract the number ;) Given your LHS of :=, it seems like you need to study ?:= and an introductory data.table text. — Henrik
– Henrik, Commented Dec 17, 2018 at 21:24

jon · Accepted Answer · 2018-12-17 21:27:29Z

1

You can use dplyr on a data.table object, which evaluates each value individually as opposed to just the first row. For example:

library(dplyr)
library(data.table)

mydt<-data.table(name = c('ao','bo'), b = c(2,2), c = c("1 GiB", "1.4 GiB"))
mydt %>% 
  mutate(d = as.numeric(gsub(" GiB","",c)))

  name b       c   d
1   ao 2   1 GiB 1.0
2   bo 2 1.4 GiB 1.4

answered Dec 17, 2018 at 21:27

jon

3681 silver badge12 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

PT2018 · Accepted Answer · 2018-12-18 18:40:42Z

0

mydt <- data.table(name = c("ao","bo"),
                   b = c(2,2),
                   c = c("1 GiB", "1.4 Gib"))

We can take several approaches.

Using tidyverse:

mydt %>% mutate(c = parse_number(c))

Or,

mydt %>% separate(col = c,into = "c",sep = " ",convert = T)

Using data.table approach

mydt[,
     c := tstrsplit(c," ", fixed = TRUE, keep = 1L)]

which gives output with column c as character.

To get c as a numeric output:

mydt[,
     c := tstrsplit(c," ", fixed = TRUE, keep = 1L)
     ][,
       c := as.numeric(c)]

Your original code can be replaced by below code:

mydt[, c :=substr(c,0,str_start(c," "))]

answered Dec 18, 2018 at 18:40

PT2018

315 bronze badges

Collectives™ on Stack Overflow

Modify value of column in data.table without loop

2 Answers 2

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related