rep_n_stack: Replicate and stack columns of a data frame

Description

Reshape a data frame by stacking two or more columns into one and adding a factor, while replicating the remaining columns and stacking them to match the number of rows

Usage

rep_n_stack(data,to.stack,stack.names=NULL)

Arguments

data

A data frame.

to.stack

Which columns are to be stacked together (see Details).

stack.names

Names for the new factor and stacked column.

Value

The reshaped data frame.

Details

rep_n_stack takes two or more specified columns in a data frame and "stacks" them into a single column. It also creates a new factor composed of the replicated names of the columns that identifies from which column each value came. The remaining columns in the data frame are replicated to match the new number of rows.

If to.stack is a matrix of names or column numbers, rep_n_stack will stack each row into two new columns, allowing multiple related sets of values to be stacked in one operation.

A matrix or data frame of values can now be stacked so that the values can be displayed by a function like barNest in the plotrix package.

Examples

Run this code

# NOT RUN {
 wide.data<-data.frame(ID=1:10,Glup=sample(c("Montic","Subtic"),10,TRUE),
  Flimit1=runif(10,1,2),Flimit2=runif(10,1.5,2.5),Flimit3=runif(10,1.2,3),
  Glimit1=rnorm(10,mean=5),Glimit2=rnorm(10,mean=4),Glimit3=rnorm(10,mean=4.5))
 # first just stack one set of related measures
 rep_n_stack(wide.data[,1:5],to.stack=c("Flimit1","Flimit2","Flimit3"))
 # now stack two sets of related measures and pass names for the stacks
 rep_n_stack(wide.data,to.stack=matrix(3:8,nrow=2,byrow=TRUE),
  stack.names=c("Limit_F","Value_F","Limit_G","Value_G"))
 # finally stack a matrix of means into a single column with the
 # row and column names becoming "factor" variables
 meanmat<-matrix(runif(16,10,20),nrow=4)
 rownames(meanmat)<-c("Plunderers","Storers","Refusers","Jokers")
 colnames(meanmat)<-c("Week1","Week2","Week3","Week4")
 rep_n_stack(meanmat,to.stack=1:4,
  stack.names=c("Returns","Occasion","Strategy"))
# }