Reputation: 15
I am having issues returning output from a function I created in R when I use it in a loop. I am trying to combine the output form multiple MCMC models into one R object.
The function:
get_scrUN_output <- function(filename){
out <- filename
nam<-c("sigma","lam0","psi", "N")
nam<-match(nam,dimnames(out[[1]]$sims)[[2]])
out.lst<-mcmc.list(
as.mcmc(out[[1]]$sims[200001:300000,nam]),
as.mcmc(out[[2]]$sims[200001:300000,nam]),
as.mcmc(out[[3]]$sims[200001:300000,nam]))
s <- summary(out.lst)
gd <- gelman.diag(out.lst,multivariate = FALSE)
output_table <- rbind(as.data.frame(t(s$statistics)),
as.data.frame(t(s$quantiles)),
as.data.frame(t(gd$psrf)))
return(output_table) }
The code I use to create a list of RData mcmc outputs to run through the function:
scrUN.ET <- list.files(getwd(),"out.*ET.RData")
scrUN.lst <- as.vector(substring(scrUN.ET,1))
scrUN.lst <- str_sub(scrUN.lst, 1, str_length(scrUN.lst)-3)
>scrUN.lst
[1] "BBout11FL" "BBout11TL" "BBout12TL" "BBout13FL" "BBout13TL"
When I use the function on an individual output file, it works:
get_scrUN_output(BBout11FL)
sigma lam0 psi N
Mean 130.43594323 14.5319368 0.3361405211 335.8042733
SD 7.28386725 9.7311139 0.2743725813 274.6828277
Naive SE 0.01329846 0.0177665 0.0005009335 0.5014999
Time-series SE 1.28032869 1.3886577 0.0360607870 36.5692414
2.5% 118.37718370 0.6129902 0.0300165600 30.0000000
25% 124.29743884 5.7535456 0.0958156210 95.0000000
50% 130.40628214 15.1264454 0.2426328827 242.0000000
75% 135.99836262 19.9685209 0.5403864215 541.0000000
97.5% 145.11615201 34.9438198 0.9298185748 930.0000000
Point est. 1.59559993 4.4590599 1.0677998255 1.0678381
Upper C.I. 2.56854388 9.5792520 1.2186078069 1.2186933
But when I try to run all output files through the function using a loop I get a NULL output.
scrUN.output <- rbind(
for (i in seq_along(scrUN.lst)){
get_scrUN_output(get(scrUN.lst[i]))
}
)
>scrUN.output
NULL
Thanks!
Upvotes: 0
Views: 124
Reputation: 4795
The reason is you're rbind
-ing nothing.
Here's a simplified example demonstrating what your code above is doing –– the for loop isn't assigning anything to a variable which is why you're getting NULL
at the end.
xx <- rbind(
for(i in c(1,2)){
i
}
)
print(xx) # NULL
Try this instead:
scrUN.output <- list() # initialize a list
for (i in seq_along(scrUN.lst)){
# update the list contents
scrUN.output[[i]] <- get_scrUN_output(get(scrUN.lst[i]))
}
# finally, rbind eveything
scrUN.output <- do.call(rbind, scrUN.output)
Or better yet, use lapply
:
scrUN.output <- lapply(scrUN.lst, get_scrUN_output)
scrUN.output <- do.call(rbind, scrUN.output)
Upvotes: 2
Reputation: 561
I think this is what you're asking for. This is an edit of the final code section. Your were using rbind on nothing since nothing is being returned by the for-loop.
scrUN.output <- lapply(scrUN.lst, function(i) get_scrUN_output(get(i)))
scrUN.output <- do.call(rbind, scrUN.output)
scrUN.output
Upvotes: 0