creating a new variable that combines several others using if else

Question

I am using a large data.frame of health conditions and outcomes, I wish to combine 10 health conditions into a single condition, If the patient has either a, or b, c, or d, etc. then the condition would be condition one. I am trying to code it like this:

      dataset$one <-  ifelse(dataset, (dataset$a == 1)|
                            (dataset$b == 1)|
                            (dataset$c  == 1)|
                            (dataset$d  == 1),  1, 0)

This seems to work for the first condition, but not when I add conditions. Perhaps R does not allow multiple or statements? Any suggestions?

You could try using `ifelse(dataset, sum(dataset[c('a','b','c','d')] == 1) >= 1, 1, 0)` — smanski, Mar 24 '18 at 17:38

G. Grothendieck · Answer 1 · 2018-11-12T16:37:46.553

3

Assuming that dataset is a data frame, define the column names, cols, and then apply any across each row of dataset[cols] == 1 like this. Add zero to convert the result from logical to numeric:

cols <- c("a", "b", "c", "d")
dataset$one <- apply(dataset[cols] == 1, 1, any) + 0

Notes

If the columns have NA values that you wish to exclude then add the na.rm = TRUE argument:
```
dataset$one <- apply(dataset[cols] == 1, 1, any, na.rm = TRUE) + 0
```
The Rfast package has rowAny which could be used if you don't need na.rm:
```
library(Rfast)
dataset$one <- rowAny(dataset[cols] == 1) + 0
```

edited Nov 12 '18 at 16:37

answered Mar 24 '18 at 17:53

G. Grothendieck

254,981
17
203
341

Rfast doesn't have a function called "anyRows"... – Manos Papadakis Aug 29 '18 at 17:18
2

Have changed it to `rowAny` – G. Grothendieck Aug 29 '18 at 17:57

score 2 · Answer 2 · answered Mar 24 '18 at 17:38

2

We can use Reduce with |

dataset$one <- as.integer(Reduce(`|`, lapply(dataset[c('a', 'b', 'c', 'd')], `==`, 1))

Or another option is rowSums

dataset$one <- as.integer(rowSums(dataset[c('a', 'b', 'c', 'd')] == 1) > 0)

answered Mar 24 '18 at 17:38

akrun

874,273
37
540
662

creating a new variable that combines several others using if else

2 Answers2

Notes