Name assignment in for loop

Tena koe Toby

Try creating a list and storing each model as an element of the list.  For example,

lmModels <- vector('list', nrow(dat2))
for (i in 1:nrow(dat2))
{
  lmModels[[i]] <- etc
}

Generally, I add a line after creating the list to name the elements and then address the list using these names as I find that less error prone and more informative.

HTH ....

Peter Alspach

-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On Behalf Of nrm2010
Sent: Friday, 2 November 2012 10:16 a.m.
To: r-help at r-project.org
Subject: [R] Name assignment in for loop



Dear helpeRs-

I'm using a for loop to create a series of models.
I'm trying to assign a name to each model created, using the loop index.  The loop gets stuck at the name  of the model, giving the error "target of assignment expands to non-language object".  The linear model runs without error; only the name is problematic.

Here is the current loop syntax.  The use of dat and dat2 is not an error.  I'm pulling data from 2 sources for the model.

for (i in 1:dim(dat2)[[1]]) {
assign("modelb",i) <- lm(log(dat$flux) ~ dat$Tsoil_flux, subset = dat$chamber == dat2$chamber[i] & dat$year == dat2$year[i] & dat$doy >= dat2$day1[i] & dat$doy <= dat2$day2[i]) dat2$coef[i] <- coef(assign("modelb",i, sep = ""))[[2]] dat2$Rsq[i] <- summary(assign("modelb",i, sep = ""))[[9]] }

I have also tried
assign("modelb",1:i) #following the ?assign example paste("modelb", i, sep = "") <- (...) assign(paste("modelb", i, sep = "")) <- (...) assign(paste("modelb", i, sep - ""), put linear model here) They all generate the same error message.


dim(dat2)[[1]] is 29

dim(dat2)[[1]]

[1] 29

class(dim(dat2)[[1]])

[1] "integer"

I have not included data because the problem is with the naming syntax;  no data are involved except for the number 29.

Given this approach, rather than the sapply() approach, what is the correct  syntax for naming each model in the sequence?

Thank you in advance.

Toby Gass