slice.tensor {tensorA} | R Documentation |
Indexing of tensors allows beside the ordinary selection of ranges of indices the renaming of indices. The functions are mainly here to keep the the tensor property of the results.
slice.tensor(X,i,what,drop=FALSE) ## Methods for class tensor # X[...,drop=TRUE] # X[...,drop=TRUE] <- value # X[[...,drop=TRUE]] # X[[...,drop=TRUE]] <- value
X |
A tensor |
i |
an index given as number or character |
what |
levels of the index, a number or a character from dimnames |
drop |
a boolean, if true, indices with only a single level are removed |
... |
arguments of the form name= indices, and for the
[[ ]] functions it also allowed to give names from the
corresponding dimnames name=c("a","b") to select indices by
names or name=~newname to rename dimensions, the first use
makes a usual array access in the given dimension, where [[ ]]
only supports a single index, while [ ] allows vectors. The
other type changes the names. |
a new tensor with dimensions renamed or individual levels selected
K. Gerald van den Boogaart
A <- to.tensor(1:20,c(A=2,B=2,C=5)) A[C=1] A[C=1:3] A[[B=~b]] # renaming dimensions A[[B=~b,A=~aaa]] A[[B=~b,A=~aaa,aaa=1]] A[[A=1,B=~gamma]][C=1:2] A