Permutation of dimensions, or generalized transpose.

`a` |
Object to be transposed. |

`perm` |
Integer vector giving the permutation. |

`...` |
Additional argument list. |

The generalised transposition here is equivalent to the corresponding
permutation of the columns of the data frame or matrix obtained with
`as.data.frame`

or `as.matrix`

. The order of the rows of the
data frame or matrix is left unchanged, and is determined by the
`index`

slot of the object. Thus if a vector or matrix of
response, say `Y`

, has its elements or rows aligned with the
design this property remains after the transposition, see
**Examples**. It must be kept in mind that this "design" order
can not be guessed when only the factors and their levels are known.

An object with the same class as `a`

but with
permuted dimensions.

`aperm`

S3 method from the
**base** package.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 | ```
myGDa <- Grid(nlevels = c("X" = 2, "Y" = 3, "Z" = 4))
myGDb <- aperm(myGDa, perm = c(3, 2, 1))
## evaluation of a function on the permuted array
myGD1 <- Grid(nlevels = c("X" = 2, "Y" = 3, "Z" = 4))
myPerm <- c(3, 2, 1)
myFun <- function(vec){ sin(vec["X"]) + vec["Y"] - vec["Z"]^2 }
## 'f1' contains the value of the function in the order of 'myGD1'
f1 <- apply_Grid(myGD1, fun = myFun)
## 'f2' contains the value of the function in the order of 'myGD2'
myGD2 <- aperm(myGD1, perm = myPerm)
f2 <- apply_Grid(myGD2, fun = myFun)
## note that 'as.matrix' sorts the observations in the index order
XYZ1 <- as.matrix(myGD1)
XYZ2 <- as.matrix(myGD2)
## check
f = apply(XYZ1, 1, myFun)
cbind(XYZ1, f1 = f1, XYZ2, f2 = f2, f = f)
eps <- sqrt(.Machine$double.eps)
all(abs(f1 - f) < eps) ## should be TRUE
all(abs(f2 - f) < eps) ## should be TRUE
``` |

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.

All documentation is copyright its authors; we didn't write any of that.