Creating a factor from a numeric column vector in R

Question

I have a numeric column of weights (df$weight0) in a dataframe df. I want to create a new column df$weight1 which is a factor based on the values in df$weight0.

If the value in df$weight0 is less than or equal to 170, the corresponding value in df$weight1 should be 1 but if the value in df$weight0 is greater than 170,the corresponding value in df$weight1 should be 2.

The code below is what I have tried but it gives a single value not a vector.

  if (i<=170){
    i==1
  }else{
    i==2
  }
}

df$weight1 <- (df$weight0>170)+1

PKumar
– PKumar

2020-02-26 12:16:51 +00:00
Commented Feb 26, 2020 at 12:16 — PKumar
– PKumar, Commented Feb 26, 2020 at 12:16

Gwang-Jin Kim · Accepted Answer · 2020-02-26 12:37:22Z

1

Use `cut()` for discretizizing continuous variables by intervals

For such kind of interval categorization, there is the very useful function cut.

nums <- nums <- runif(100, min=0, max=300) # n = 100 random numbers between 0 and 300
factorized_num <- cut(nums, c(-Inf, 170, +Inf))
# you can name the categories as you want:
levels(factorized_num) <- c(1, 2) # first interval 1, next interval 2
# with include.lowest=TRUE or FALSE you can determine whether lower limit is <= or <

Or use `Vectorize()` to vectorize non-vectorized functions

# define for one case:
categorize <- function(i) if (i<=170) 1 else 2
# then vectorize it:
categorize <- Vectorize(categorize)

Now you can use it:

categories <- categorize(nums)
head(categories) ## 1 2 1 1 ...

I prefer this - out of my experience - much more over ifelse() because you have full control over the single case.

edited Feb 26, 2020 at 12:37

answered Feb 26, 2020 at 12:26

Gwang-Jin Kim

11k20 silver badges39 bronze badges

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

2 Comments

Ama Owusu-Darko Over a year ago

Thank you all. I tried with the cut method wcgs$w8t<-cut(wcgs$weight0,breaks = c(78,170,320), labels = c("<= 170 lbs", "> 170 lbs"), include.lowest=T)

Gwang-Jin Kim Over a year ago

nice! The best is you include/name all limits (and -Inf and +Inf are thereby useful if there are no limits).

SebSta · Accepted Answer · 2020-02-26 12:14:25Z

1

ifelse can use vectorized input:

df$weight1 <- ifelse(df$weight0<=170,1,2)

answered Feb 26, 2020 at 12:14

SebSta

5062 silver badges12 bronze badges

Comments

Daniel O · Accepted Answer · 2020-02-26 12:17:24Z

1

You were checking the value of i, not the values in your df. Also the assignment of your new columns was not implemented. Try the following.

for (i in wcgs$weight0){
  if (wcgs$weight0[i]<=170){
    wcgs$weight1[i] <- 1
  }else{
    wcgs$weight1[i] <- 2
  }
}

answered Feb 26, 2020 at 12:17

Daniel O

4,3789 silver badges21 bronze badges

1 Comment

Ama Owusu-Darko Over a year ago

Thanks. So I use wcgs$weight1[i] to change the value at that index

massisenergy · Accepted Answer · 2020-02-26 12:31:08Z

0

Using case_when from the package dplyr:

library(dplyr)
df %>% mutate(df$weight1 = case_when(df$weight0 =< 170 ~ 1,
                                     df$weight0 > 170 ~ 2))

case_when() helps to formulate if-else construct and mutate() modifies or adds columns.

edited Feb 26, 2020 at 12:31

answered Feb 26, 2020 at 12:20

massisenergy

1,8843 gold badges16 silver badges30 bronze badges

Collectives™ on Stack Overflow

Creating a factor from a numeric column vector in R

4 Answers 4

Use `cut()` for discretizizing continuous variables by intervals

Or use `Vectorize()` to vectorize non-vectorized functions

2 Comments

Comments

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

Use cut() for discretizizing continuous variables by intervals

Or use Vectorize() to vectorize non-vectorized functions

2 Comments

Comments

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related

Use `cut()` for discretizizing continuous variables by intervals

Or use `Vectorize()` to vectorize non-vectorized functions