R-Cookbook.com Delicious, statistical-computing recipes

The function and documentation is listed after the jump.

You may need to stretch out your browser window to see the whole thing.

# Description:
#	Some data is stored in univariate or stacked form, where each row contains data
#	from one individual per group, or in the case of logitudinal data, from
#	one observation per subject.  Sometimes, however, it is more convenient
#	when data is in a multivariate form, where each row contains data from all
#	individuals per group, or in the case of logitudinal data, from all
#	observations per subject.
# Usage:
#	make.multiv(univ, group_column, individual_column, variable_columns, sep=".")
# Arguments:
#	univ - The univariate data set; must have the same number of individuals per
#		group, for every group
#	group_column	- The column number (index) or name of the grouping id column
#	individual_column - The column number or name of ids of the individuals that
#		belong to the groups
#	variable_columns - A list of column numbers or names of the variables measured
#		for each individual in each group
#	sep - The seperator for the new variable column names that goes between the
#		variable name and the individual id
# Value:
#	Returns a dataset with all of the variables for each individual in each group
#		on a single row
# Author(s):
#	Jeff; updated at http://www.r-cookbook.com/node/59

make.multiv <- function(univ, group_column, individual_column,
					variable_columns, sep="."){
	.by <- by(univ[variable_columns], univ[group_column], unlist)
	.vs <- t(sapply(.by, c))
	.id <- data.frame(unique(univ[group_column]), row.names=NULL)
	.varpart <- rep(names(univ[variable_columns]),
					each=length(unlist(unique(univ[individual_column]))))
	.iidpart <- rep(unlist(unique(univ[individual_column])),
					length(variable_columns))
	.names <- paste(.varpart, .iidpart, sep=sep)
	multiv <- cbind(.id, .vs)
	names(multiv) <- c(names(univ[group_column]), .names)
	multiv
}

# Examples

univ1 <- read.table(
	textConnection("
		FID TID TRAIT1 TRAIT2
		1 1 4 5
		1 2 9 10
		2 1 2 3
		2 2 2 4
		3 1 11 12
		3 2 0 1
	"), header=TRUE)

make.multiv(univ1, 1, 2, c(3, 4), sep=".")
make.multiv(univ1, "FID", "TID", c("TRAIT1", "TRAIT2"), sep=".")

#   FID TRAIT1.1 TRAIT1.2 TRAIT2.1 TRAIT2.2
# 1   1        4        9        5       10
# 2   2        2        2        3        4
# 3   3       11        0       12        1

univ2 <- read.table(
	textConnection("
		FID TID V1 V2 V3
		1 1 4 5 6
		1 2 9 10 5
		2 1 2 3 4
		2 2 2 4 4
		3 1 11 12 1
		3 2 0 1 0
	"), header=TRUE)

make.multiv(univ2, 1, 2, c(3:5), sep="-TWIN")
make.multiv(univ2, "FID", "TID", c("V1", "V2", "V3"), sep="-TWIN")

#   FID V1-TWIN1 V1-TWIN2 V2-TWIN1 V2-TWIN2 V3-TWIN1 V3-TWIN2
# 1   1        4        9        5       10        6        5
# 2   2        2        2        3        4        4        4
# 3   3       11        0       12        1        1        0