Problems with loop/repeat in R - r

Question

I need to execute this code many times in order to get 45 different matrices at the end: mat[j], j=1:45.
Not sure how to use "for-loop" to achieve that, will be grateful for any tips.
Data files are stored here, year-by-year https://intl-atlas-downloads.s3.amazonaws.com/index.html
library(readstata13)
library(diverse)
library(plyr)
for (j in 1:45) {
dat <- read.dta13(file.choose())
data = aggregate(dat$export_value, by = list(dat$exporter,dat$commoditycode), FUN = sum)
colnames(data) = c("land","product","value")
dt = split(data, f = data$product)
land = as.data.frame(sort(unique(data[, 1])))
nds = seq(1, nrow(land), by = 1)
texmat = cbind(nds, land)
colnames(texmat) = c("num", "land")
for (i in 1:length(unique(data[, 2]))) {
(join(texmat, dt[[i]], by = "land", type = "left")$value)
}
mt = sapply(1:length(unique(data[, 2])), function(i) join(texmat, dt[[i]], by = "land", type = "left")$value)
colnames(mt) = unique(data[, 2])
rownames(mt) = sort(unique(data[, 1]))
mt[is.na(mt)] = 0
rcamat=values(mt, category_row = FALSE, norm = "rca",filter = 1, binary = TRUE)
rcamat[is.na(rcamat)] = 0
tmat = rcamat[rowSums(rcamat) != 0, , drop = TRUE]
mat = t(tmat)
}

Answer 1

It looks like you're almost there with the for loop. You just need to add 2 concepts:
1) Creating a list of matrices to read at the start. A construction like:
filenames <- paste0('H0_',1995:2016,'.dta')
filenames <- c(filenames,paste0('S2_final_',1962:2016,'.dta'))
that creates a vector of the files you want to read will allow you to replace file.choose with something like the following (inside the loop):
dat <- read.dta13(paste0('/path/to/directory/with/files/',filenames[i]))
This way you can grab a new file with each loop iteration.
2) Storing the output matrices at the end of the loop. You can do this either by putting them all in a list, or by using assign to create a collection of objects. I prefer the list approach:
#before the for loop initialize a NULL list:
mats <- NULL
#at the end of the loop, (after mat = t(tmat) but before the close bracket) add this line to add it to the list
mats[[i]] <- mat
This will create a list mats with mats[[1]] holding the first matrix, mats[[2]] holding the second, and so on.
You could alternatively create a bunch of objects like so:
#at the end of the for loop add
assign(paste0('mat_',i),mat)
Which will create mat_1, mat_2, and so on as separate objects. A full implementation would look something like this:
library(readstata13)
library(diverse)
library(plyr)
setwd('/path/to/files/')
filenames <- paste0('H0_',1995:2016,'.dta')
filenames <- c(filenames,paste0('S2_final_',1962:2016,'.dta'))
#you'll have to prune this to the files you actually want, as this list is more than 45
finished_matrices <- NULL
for (j in 1:45) {
dat <- read.dta13(filenames[i]) #pickup
data = aggregate(dat$export_value, by = list(dat$exporter,dat$commoditycode), FUN = sum)
colnames(data) = c("land","product","value")
dt = split(data, f = data$product)
land = as.data.frame(sort(unique(data[, 1])))
nds = seq(1, nrow(land), by = 1)
texmat = cbind(nds, land)
colnames(texmat) = c("num", "land")
for (i in 1:length(unique(data[, 2]))) {
(join(texmat, dt[[i]], by = "land", type = "left")$value)
}
mt = sapply(1:length(unique(data[, 2])), function(i) join(texmat, dt[[i]], by = "land", type = "left")$value)
colnames(mt) = unique(data[, 2])
rownames(mt) = sort(unique(data[, 1]))
mt[is.na(mt)] = 0
rcamat=values(mt, category_row = FALSE, norm = "rca",filter = 1, binary = TRUE)
rcamat[is.na(rcamat)] = 0
tmat = rcamat[rowSums(rcamat) != 0, , drop = TRUE]
mat = t(tmat)
finished_matrices[[i]] <- mat
}