R: make 2 subset vectors so that values are different index-wise, and also different across each vector

问题

Following up on this question, I want to do something similar, but this time I have one more requirement.

I want to make 2 vectors subsetting from the same data.

I need replace to be set to FALSE because I need all values to be different across a, and all values to be different across b.

Apart from that, values cannot be the same in a and b for the same index position.

Note that sampling vector v is always fixed, as is the sample length l.

Doing the following, I only fulfil one criterium (values across a and values across b are different, but still values in the same index between a and b can be identical)

> set.seed(1)
> v <- 1:15
> l <- 10
> a <- sample(v, l, replace=F)
> b <- sample(v, l, replace=F)
> a
 [1]  4  6  8 11  3  9 10 14  5  1
> b
 [1]  4  3  9  5 13 12  7  8 14 10
> a==b
 [1]  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE

Doing the following (answer from the previous question), I only fulfil the other criterium (values in the same index are not identical, but there can be identical values across a or b).

> ab <- split(replicate(10, sample(15,2)), seq(2))
> a <- ab[[1]]
> b <- ab[[2]]
> a
 [1] 14  7 10  8  2  6  6  8 13 12
> b
 [1]  5  5  4 11 13 12  5 13  6 14
> duplicated(a)
 [1] FALSE FALSE FALSE FALSE FALSE FALSE  TRUE  TRUE FALSE FALSE
> duplicated(b)
 [1] FALSE  TRUE FALSE FALSE FALSE FALSE  TRUE  TRUE FALSE FALSE

Any help trying to collapse both approaches? Thanks!

回答1:

Edited with a while loop to do the resampling until it is correct (with no duplicates).

Got it... What I did was check which values in b are equal to a for the same index, and which are not equal.

For those that are equal, I resample from a vector equal to v but without the already "accepted" values in b (the ones not equal to a for the same index).

Like this:

> set.seed(6)
> v <- 1:15
> l <- 10
> a <- sample(v, l, replace=F)
> b <- sample(v, l, replace=F)
> a
 [1] 10 14  4  5  9 15 11  7 13  1
> b
 [1] 10 13  2  4  9  3  5  6 14 11
> a==b
 [1]  TRUE FALSE FALSE FALSE  TRUE FALSE FALSE FALSE FALSE FALSE


> if (any(a==b)==TRUE) {
+   b0 <- b[which(a==b)]
+   b2 <- b[which(a!=b)]
+   vnew <- v[which(!(v %in% b2))]
+   b0 <- sapply(b0, function(x) sample(vnew[vnew != x], 1))
+   while (any(duplicated(b0))==TRUE){
+     b0 <- sapply(b0, function(x) sample(vnew[vnew != x], 1))
+   }
+   b[which(a==b)] <- b0
+ }


> a
 [1] 10 14  4  5  9 15 11  7 13  1
> b
 [1] 15 13  2  4 12  3  5  6 14 11
> a==b
 [1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
> duplicated(b)
 [1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE

来源：https://stackoverflow.com/questions/58375977/r-make-2-subset-vectors-so-that-values-are-different-index-wise-and-also-diffe

标签

indexing

subset

element