[R] create an R object in a loop
Ivan Calandra
ivan.calandra at uni-hamburg.de
Fri Jan 29 17:39:21 CET 2010
Hi everybody,
To run some statistical tests from the package WRS (from Rand R Wilcox),
I need to store my data in a list, which fac2list() from this package
does very well.
But I would like to do it in a loop for each numerical variable. It
would be easier!
For now, I have the loop with the extraction and storage into a list.
The code is below.
-----
# Defines variables
file <- ssfamed #data file object to be used
spec <- "cotau" #species selection
bone <- "tx" #tooth row selection
cat <- 3 #column indexe to be used as
categorical variable
seq.num <- c(seq(7,21,1)) #column indexes to be used as numerical
variables
# Select data for species and tooth row
select <- file[file$SPECSHOR==spec & file$BONE==bone, ]
#Separate the data from each numeric variable in seq.num (select[,k])
into groups from levels in select[cat] and store into list mode.
for(i in 1:length(seq.num)) {
k <- seq.num[i]
name.num <- names(file)[k]
TO_POS_Asfc.median_cotautx <- fac2list(select[,k], select[cat])
names(TO_POS_Asfc.median_cotautx) <- levels(factor(select[[cat]]))
}
-----
What I want to do is, instead of giving manually the name of the list
(here TO_POS_Asfc.median_cotautx), I would like the name of the list to
be created in the loop too, so that it looks like:
select[[cat]]_name.num_specbone. In the code it would look like:
-----
select[[cat]]_name.num_specbone <- fac2list(select[,k], select[cat])
#create the list from the values of these variables at each iteration of
the loop
names(select[[cat]]_name.num_specbone) <- levels(factor(select[[cat]]))
-----
I thought about using paste(), but I cannot create an object like this.
In my small R life, I've always stored data into pre-defined objects, I
have no idea how to create an object automatically within a loop.
Is my question clear? I hope so. I thought there might be a function
that would create an object where the name would come from a string
stored in an other object.
Why do I want that? Simply because I want to run several statistical
tests on each list and I would therefore like to have standardized names
for the data objects.
There might be easier ways to do what I want to, I'm open to all
suggestions but I would prefer to stay close from my idea (I would
understand better!)
Thanks in advance for your help.
Have a nice weekend
Ivan
More information about the R-help
mailing list