amult: Generalized array multiplication.

In garray: Generalized Array Arithmetic for Ragged Arrays with Named Margins

Description

Default to Einstein summation convention, without explicitly subscripts.

Usage

amult(X, Y, FUN = "*", SUM = "sum", BY = NULL, MoreArgs = NULL,
  ..., SIMPLIFY = TRUE, VECTORIZED = TRUE)

X %X% Y

Arguments

X, Y	Generalized arrays that can be multiplied.
FUN	The 'multiply' function.
SUM	The 'reduce' function.
BY	margins excluded from summary by SUM.
MoreArgs, SIMPLIFY, VECTORIZED	Argument used by 'amap()'.
...	Argument used by 'areduce()'.

Details

Margins shared by X and Y are parallelly mapped by FUN, and then reduced by SUM (inner product like %*%); margins in BY and shared by X and Y are simply mapped by FUN but excluded from reducing (parallel product like *); other margins are extended repeatly (outer product like %o%). Shared margins not to be mapped have to be renamed (like outer product). For special FUN and SUM, fast algorithms are implemented.

Examples

a <- garray(1:24, c(4,6), list(X=LETTERS[1:4], Y=letters[1:6]),
	sdim=list(XX=c(x1=3,x2=1), YY=c(y1=1,y2=2)))
b <- garray(1:20, c(Z=5, X=4))
c <- garray(1:120, c(X=4,Y=6,Z=5))
m1 <- amult(a, b)
m2 <- amult(a, b, `*`, sum)
m3 <- amult(b, a)
all.equal(m1, m2)
all.equal(m1, m3)
all.equal(m1, t(m3))
n1 <- amult(a, c, `*`, sum)
n2 <- a%X%c
all.equal(n1, n2)
amult(garray(1:5,margins="I"), garray(1:8,margins="J"))
amult(garray(1:8,c(I=2,J=4)), garray(1:9,c(K=3,L=3)))

garray documentation built on May 1, 2019, 8:47 p.m.