interaction: Compute the Interaction of Several Factors
Description
This is a replacement for the builtin function interaction that adds
the parameters sep and left.
Usage
interaction(..., drop=FALSE, sep=".", left=FALSE)
Arguments
...
the arguments to interaction can be either a data frame containing
all the factors to be used
all the individual factors.
It will not understand a combination of factors and designs as arguments;
you have to pick one form or the other.
drop
if TRUE the levels of the new factor not represented in the data are dropped.
sep
the separator in creating the level descriptors.
left
set to TRUE to left-justify factor levels when constructing label strings
Value
a new factor, whose levels are all possible combinations of the factors
supplied as arguments. If drop =TRUE, only the levels represented in the
new factor are retained.
Details
This is a slight modification of the S-supplied function. The sep argument
has been added, a sort step has been added to properly sort the levels of
the created variable, and the left parameter is added.