Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

How to export multiple function or packages in foreach loop in "R"

I'm trying to decrease run time of my code by using doParallel package in R.

I'm calling a function awareRateSIR that some extra packages are used in the body of this function. I get some error like

could not find function "vcount" and..

I know vcount is a function of package igraph that is used in awareRateSIR) but it's not the only one. How can I solve this problem? I've thought I should pass all packages name that is used in my function awareRateSIR but I don't know how I cant export multiple function in foreach or how I can export multiple package name.

This is my code:

tp<-foreach(i=1:iter,  .inorder = FALSE, .export = "awareRateSIR",
          .packages = "igraph", .packages="doParallel")%dopar%{
tp <- awareRateSIR(graphContact, graphCom,state)
return(tp)
}

if I don't pass these packages I will get error states some function are unknown if I pass all the packages I will get error:

Error in foreach(i = 1:iter, .inorder = FALSE, .export = "awareRateSIR", : formal argument ".packages" matched by multiple actual arguments"

thanks in advance

like image 350
ElhamMotamedi Avatar asked Aug 18 '17 07:08

ElhamMotamedi


2 Answers

You should pass all the packages required, using c function, as below :

tp<-foreach(i=1:iter,  .inorder = FALSE, .export = "awareRateSIR",
          .packages = c("igraph", "doParallel"))%dopar%{
tp <- awareRateSIR(graphContact, graphCom,state)
return(tp)
}
like image 63
Kumar Manglam Avatar answered Nov 04 '22 14:11

Kumar Manglam


Usually, you don't need to export any variable, function or package when using foreach.

For variables and functions, they just need to be in the same environment as the foreach statement (the better is to make a function where you pass only what you need).

When functions are required from other packages, you just need to use package::function, like inside packages.


E.g. you can do

library(doParallel)
registerDoParallel(cl <- makeCluster(3))

df <- iris

foreach(spec = unique(iris$Species)) %dopar% {
  dplyr::filter(df, Species == !!spec)
}

stopCluster(cl)
like image 22
F. Privé Avatar answered Nov 04 '22 15:11

F. Privé