[R] randomness using runif

[Duncan Murdoch]

On 23/07/2010 11:16 AM, Brigid Mooney wrote:
> I'm  working on a problem where I'm introducing random error and have
> been using the built in function runif to provide that random error.
> However, I realized that I seem to be getting some unexpected behavior
> out of the function and was hoping someone could share some insight.
>
> I don't know the runif algorithm at all, but from the behavior I'm
> seeing, it seems that whenever I open a new R console, the function
> runif gets "reset" to some initial value.  For example...
>
> In a NEW R console, enter the following:
>
> x1 <- runif(1000, -1, 1)
> x2 <- runif(1000, -1, 1)
>
> x1[1:5]
> x2[1:5]
>
> objectsToSave <- c("x1", "x2")
> filename <- "C:\\Documents\\x1x2file.Rdata"
>
> save(list=objectsToSave, file=filename, envir = parent.frame())
>
>
> Then in a different NEW R console, enter this:
>
> x3 <- runif(1000, -1, 1)
> x4 <- runif(1000, -1, 1)
>
> x3[1:5]
> x4[1:5]
> # For me, the values look identical to x1 and x2, but let's check by
> loading the x1x2 file and comparing them directly...
>
> filename <- "C:\\Documents\\x1x2file.Rdata"
>
> load(filename)
>
> sum(x1==x3)
> sum(x2==x4)
>
>
> For my results, I get that x1=x3 for all 1000 elements in the vector,
> and x2=x4 for all 1000 elements in that vector.
>
> Does anyone have insight into what's going on here?  Am I doing
> something wrong here, or is this a quirk of the runif algorithm?  Is
> there a better function out there for seeding truly random error?
>   

You are doing something wrong:  At some point in the past, you saved 
your workspace, but more recently, you haven't been saving it.  Your 
workspace contains the key used by all the random number generators, so 
you keep restoring the same old key.

I would advise that you *never* save your workspace, and if you see the 
message:

[Previously saved workspace restored]

you treat that as an error, and try to track down where the old 
workspace came from.  Some other people would advise that you always 
save your workspace.  Both approaches will solve this problem.
> For what it's worth, here's my R version info:
>
> platform       i386-pc-mingw32
> arch           i386
> os             mingw32
> system         i386, mingw32
> status
> major          2
> minor          8.1
> year           2008
> month          12
> day            22
> svn rev        47281
> language       R
> version.string R version 2.8.1 (2008-12-22)
>   

I'd also suggest upgrading: there have been a lot of bug fixes and new 
features added since 2.8.1.

Duncan Murdoch

> Thanks for the help,
> Brigid
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>