Calling variable works outside of loop, but not inside loop - r

Question

I am looping through vars_macro. The first variable in vars_macro is c1372 (dput below). The below code works perfectly fine:
len <- 32
c1372[1:(len-z),1:1]
However when I try to call the same variable (c1372) in code below, I get an error:
Error in m[1:(len - z), 1:1] : incorrect number of dimensions
Code:
output <- list()
forecast <- list()
for(m in noquote(vars_macro)){
output[[m]] <- list() # treat output as a list-of-lists
fit[[m]] <- list() # treat fit as a list-of-lists
for(z in rev(1:6)) {
x <- m[1:(len-z),1:1]
x <- ts((x), start = c(2011, 4), frequency = 4)
y <- Macro[1:(len-z),2:2]
y <- ts((y), start = c(2011, 4), frequency = 4)
t <- Macro[(len+1-z):(len+1-z),3:10]
t <- ts((t), start = c(2019, 2), frequency = 4)
#fit model
fit[[m]][[z]] <-auto.arima(y,xreg=x,seasonal=TRUE)
output[[m]][[z]] <- forecast(fit[[m]][[z]],xreg=t)$mean
}
}
Note the code above fails on the first variable (c1372), so the issue isn't the other variables. You can confirm this by simply writing for(m in ("c1372"))
Dput:
dput(vars_macro)
c("c1372", "c5244", "c5640", "c6164", "b1372", "b5244", "b5640",
"b6164", "v1372", "v5244", "v5640", "v6164", "bv1372", "bv5244",
"bv5640", "bv6164")
dput(c1372)
structure(list(c1372 = c(1.386445329, 1.600103663, 1.906186443,
1.962067415, 2.716663882, 1.875961101, 2.086589462, 2.115101307,
2.960605275, 2.109288864, 2.730920081, 2.816577742, 4.006180002,
3.503741762, 4.162132837, 4.122407811, 5.352681171, 3.961705849,
4.773003078, 4.575654378, 5.71727247, 4.401603262, 5.204187541,
4.7354794, 5.809822373, 4.137968937, 4.881120131, 4.812274313,
6.143882981, 4.935116748, 5.95001413, 5.384694268)), row.names = c(NA,
-32L), class = "data.frame")

Answer 1

The code in the OP fails because once the line for(m in noquote(vars_macro)) executes for the first time, m is set to a single element character vector, c1372.
Therefore, x <- m[1:(len-z),1:1] fails because m is a single element character vector, not a data frame with 32 rows and one column.
In R, everything is an object, and it's important to know the types of objects one is manipulating. The mechanism to move back and forth between a character vector and an actual object are the two R functions, get() and assign().
assign() assigns a name with an object. get() retrieves an object, given a name.
If what is required is to access the c1372 data frame instead of the character vector c1372, one needs to use the get() function to get a named object.
Illustrating with the data provided in the OP:
vars_macro <- "c1372"
c1372 <- structure(list(c1372 = c(1.386445329, 1.600103663, 1.906186443,
1.962067415, 2.716663882, 1.875961101, 2.086589462, 2.115101307,
2.960605275, 2.109288864, 2.730920081, 2.816577742, 4.006180002,
3.503741762, 4.162132837, 4.122407811, 5.352681171, 3.961705849,
4.773003078, 4.575654378, 5.71727247, 4.401603262, 5.204187541,
4.7354794, 5.809822373, 4.137968937, 4.881120131, 4.812274313,
6.143882981, 4.935116748, 5.95001413, 5.384694268)), row.names = c(NA,
-32L), class = "data.frame")
len <- 32
theData <- NULL
for (m in vars_macro){
theData <- get(m)
}
# print first few rows to show that get() worked
head(theData)
...and the output:
> # print first few rows to show that get() worked
> head(theData)
c1372
1 1.386445
2 1.600104
3 1.906186
4 1.962067
5 2.716664
6 1.875961
>