Merging binary variables in R to a new variable

user_17985256 0
Viewed 71

I have three categorical variables i.e. stroke, MI and BP with values of 0 = yes and 1= No. I want to merge them to make a new variable "cvd" out of these three variables where each row with 0 gets 0 values in new cardiovascular variable. For example:

Stroke  MI  BP  CVD
0       1    1   0
1       1    1   1
1       1    0   0

I tried the following code but this is not what i want

transform(koratest, cvd=paste(stroke,MI, BP))

Can someone please help what could be the script for this?

Best,

Thank you for all the solutions. What to do if there is missing values in any of the values to be merged. I want missing values to be labelled as 1 but if there is 0 with missing value, i want cvd variable to have value of 1. For example:

 Stroke  MI  BP  CVD
0       1    1   0
1       NA   NA  1
0       NA   1   0

How could i achieve such output?

r
8 Answers

Try,

(rowSums(df) == ncol(df)) * 1
#[1] 0 1 0
user_56355800

Another way:

library(dplyr)

df <- data.frame(Stroke = c(0,1,1),
                   MI = c(1,1,1),
                   BP = c(1,1,0))

df %>% 
  rowwise() %>% 
  mutate(
    CVD = min(Stroke, MI, BP) 
  ) %>% 
  ungroup()

#> # A tibble: 3 × 4
#>   Stroke    MI    BP   CVD
#>    <dbl> <dbl> <dbl> <dbl>
#> 1      0     1     1     0
#> 2      1     1     1     1
#> 3      1     1     0     0

Created on 2022-07-11 by the reprex package (v2.0.1)

user_108583210

Try this using dplyr rowwise function

library(dplyr)

df |> rowwise() |> mutate(CVD = if(all(c_across() == 1)) 1 else 0) |> ungroup()
  • output
# A tibble: 3 × 4
# Rowwise: 
  Stroke    MI    BP   CVD
   <int> <int> <int> <dbl>
1      0     1     1     0
2      1     1     1     1
3      1     1     0     0
user_191983370

Don't know how you arrange your variables. If they are separted vectors, this should work:

Stroke = c(0,1,1)
MI = c(1,1,1)
BP = c(1,1,0)
CVD = as.numeric(Stroke & MI & BP)

If a data.frame:

df$CVD = with(df, as.numeric(Stroke & MI & BP)

Or the solutions mentioned by others.

user_81917270

Maybe this:

library(tidyverse)

Data <- data.frame(Stroke = c(0,1,1),
                   MI = c(1,1,1),
                   BP = c(1,1,0))

Data <- Data %>% 
  mutate(CVD = if_else(Stroke == 1 &MI == 1 & BP == 1, 1, 0))
user_120349720

base R option:

df$CVD <- apply(df,2, function(x) !any(0 %in% x)) + 0
df

Output:

  Stroke MI BP CVD
1      0  1  1   0
2      1  1  1   1
3      1  1  0   0
user_142827140

Using rowSums in cbind detects that dat is a data frame and creates such.

cbind(dat, CVD=+(rowSums(dat[c('Stroke', 'MI', 'BP')]) == 3))
#   Stroke MI BP CVD
# 1      0  1  1   0
# 2      1  1  1   1
# 3      1  1  0   0

If you only have these columns, it simplifies to:

cbind(dat, CVD=+(rowSums(dat) == 3))

Data:

dat <- structure(list(Stroke = c(0L, 1L, 1L), MI = c(1L, 1L, 1L), BP = c(1L, 
1L, 0L)), class = "data.frame", row.names = c(NA, -3L))
user_65740380

Another way to solves your problem:

df$CVD = with(df, pmin(Stroke, MI, BP)) 

  Stroke MI BP CVD
1      0  1  1   0
2      1  1  1   1
3      1  1  0   0

# or
library(data.table)

setDT(df)[, CVD := pmin(Stroke, MI, BP)]

# or
library(dplyr)

df = df %>% 
  mutate(CVD = pmin(Stroke, MI, BP))
user_108488980
Related