CODEWITHSUNDEEP
rfunctionif-statementconditional-statementsswap
sklal
sklal

Reputation: 175

How to swap two columns in R based on a specific condition?

I have two columns,say, a & b, and I have to swap the values of a & b if the value in column b is greater than the value in column a.

I've written a function to swap the variable if the condition is met

a     b
110   70
120   80 
80    110 


swap_if<- function (a, b,missing=NA) 
{
  if(b>a)
  {
    a = a + b
    b = a - b
    a = a - b
  }
}  


swap_if(a,b)

The output should come as :

a     b
110   70
120   80 
110   80

But I'm getting an error 

the condition has length > 1 and only the first element will be used

Upvotes: 2

Views: 1443

Answers (3)

hmhensen
hmhensen

Reputation: 3195

Gregor's solution is the best, but thought I'd add an extra.

swap_if <- function (a, b, missing = NA) {
  c <- a
  x <- ifelse(b > a, b, a)
  y <- ifelse(b <= a, b, c)
  z <- data.frame(x, y)
  return(z) }

swap_if(a, b)
    x  y
1 110 70
2 120 80
3 110 80

Upvotes: 2

Gregor Thomas
Gregor Thomas

Reputation: 145755

If it's only 2 columns, pmin and pmax are your friends here.

temp_min = pmin(df$a, df$b)
df$a = pmax(df$a, df$b)
df$b = temp_min

If it's more than 2 columns, than the apply sort scales better.

Upvotes: 6

ThomasIsCoding
ThomasIsCoding

Reputation: 101034

Assuming your input is given as 

df <- data.frame(a = c(110,120,80), b = c(70,80,110))

Then the following can give the expected output:

df_swap <- as.data.frame(t(apply(df, 1, function(v) sort(v,decreasing = T))))
colnames(df_swap) <- names(df)

such that

> df_swap
   a b
1 110 70
2 120 80
3 110 80

Upvotes: 2

Related Questions