Advertisements

Suppose my Dataframe look like:

```
w <- sample(-10:10, size =10)
x <- sample(-10:10, size =10)
y <- sample(-10:10, size =10)
z <- sample(-10:10, size =10)
df <- data.frame(w,x,y,z)
```

Now I want to calculate the amout of the difference (|w-x|) of each column to all the others.So I want it like the example down but additionally also wie the columns w_z, and x_y, y_z, y_z ….(my orignial df have some more Variables)

```
df$w_x <- abs(df$w-df$x)
df$w_y <- abs(df$w-df$x)
```

Alternativly, if there are easyer solution, its ok to have the calculation to a new df / matrix. Thanks in advance.

### >Solution :

Here is a method with `combn`

```
nm1 <- combn(names(df), 2, FUN = paste, collapse = "_")
df[nm1] <- combn(df, 2, FUN = function(x) abs(x[[1]]- x[[2]]))
```

-output

```
> df
w x y z w_x w_y w_z x_y x_z y_z
1 7 3 -2 6 4 9 1 5 3 8
2 -6 6 7 0 12 13 6 1 6 7
3 -8 0 -7 8 8 1 16 7 8 15
4 9 -9 -4 -2 18 13 11 5 7 2
5 5 -3 5 -4 8 0 9 8 1 9
6 2 -6 -8 -10 8 10 12 2 4 2
7 -5 7 2 -9 12 7 4 5 16 11
8 3 -2 4 -6 5 1 9 6 4 10
9 -10 -4 9 9 6 19 19 13 13 0
10 8 5 10 4 3 2 4 5 1 6
```