supermat {lazy.mat}

R Documentation

Declare a supermatrix with SMdim attribute

Description

Declare a supermatrix with SMdim attribute

Usage

supermat(nrows = 1, ncols = 1, initvar = 0)

Arguments

nrows	a vector of rows of each submatrix
ncols	a vector of columns of each submatrix.
initvar	initial value of the elements

Details

Note on supermatrix
A supermatrix, which sometimes is called as a partitioned matrix or a block matrix, is a matrix whose elements are matrices.

For example, a super matrix shown below consits of 9 submatrices:
.....c1...c1...c2...c1...c2...c3...c4
r1 1111 1112 1212 1113 1213 1313 1413
r2 2111 2112 2212 2113 2213 2313 2413

r3 1121 1122 1222 1123 1223 1323 1423
r4 2121 2122 2222 2123 2223 2323 2423
r5 3121 3122 3222 3123 3223 3323 3423

r6 1131 1122 1222 1123 1233 1333 1433
r7 2131 2122 2222 2123 2233 2333 2433
r8 3131 3122 3222 3123 3233 3333 3433
r9 4131 4132 4232 4123 4233 4333 4433


Using lazy.tools and lazy.mat packages, we can create the above by
SM <- demomat(c(2,3,4),c(1,2,4),supermat=1) .
The generated SM matrix has 9 rows and 7 colums and the SMdim attributes, which, in this case, is a list:
list(nrows=c(2,3,4),mcols=c(1,2,4)) .

How to define supermatrix:
1) Use demomat function and modify the submatrices later using modSM function.
2) Use supermatrix function to initializes a super matrix with the SMdim attributes and assign the submatrices later. using modSM function.
3) Manually define a supermatrix as a usual matrix and add the SMdim attributes to it.

The sub matrices can be accessed by getSM function and can be modifed by modSM function.

To transpose a supermatrix, use transp function instead since it preserves the SMdim attributes.

To print a super matrix with its submatrices being separated, use Printb function.

About SMdim attribute:
SMdim attribute of a supermatrix is a list of the form (nrows=, ncols=) where
nrows is a vector consisting of the # of rows of each submatrix,
ncols is a vector consisting of the # of columns of each submatrix.

Value

a sum(nrow) x sum(ncols) matrix with SMdim attribute: SMdim=list(nrows=nrows,ncols=ncols)

Examples

# using demomat function
SM <- demomat(c(2,3,4),c(1,2,4),supermat=1)
Printb(SM)
getSM(SM,2,3)

# transposition
SMt <- transp(SM)
Printb(SMt)
getSM(SMt,2,3)

# using surmat funcgtion and modSM function
( SM <- supermat( c(2,3,4),c(1,2,4) ) )
for( i in 1:3 )
 for( j in 1:3 )
  modSM(SM, i,j) <- 10*i+j
Printb(SM)

# using surmat funcgtion and getSM function with index option
( SM <- supermat( c(2,3),c(3,4,1) ) )
for( i in 1:2 )
 for( j in 1:3 )
  SM[ getSM(SM,i,j,1)$rows, getSM(SM,i,j,1)$cols ] <- 10*i+j
Printb(SM)

# adding the SMdim attributes later
SM <- matrix( 0, sum(c(2,3)), sum(c(3,4)) )
Printb(SM)
attributes(SM)$SMdim <- list(nrows=c(2,3),ncols=c(3,4))
Printb(SM)

# horizontally stacking matrices
Mh="
    c1     c1  c2       c1  c2  c3  c4
r1 111    112 122      113 123 133 143
r2 211    212 222      213 223 233 243
r3 311    312 322      313 323 333 343
r4 411    412 422      413 423 433 443
r5 511    512 522      513 523 533 543
";
Mh=cards( Mh, header=1 )
attributes(Mh)$SMdim=list(nrows=5, ncols=c(1,2,4))
Printb(Mh, sep="  ")

---