R help - Mar 2012 - scalar assignment within a vector within function

Hello,

I'm trying to create a vector of r^2 values for using a function which I
will run in a "for" loop. Example:

per<-rnorm(100,.5,.2)^2
x<-rnorm(100,10,5)
y<-rnorm(100,20,5)
fr<-data.frame(x,y,per)

test<-rep(0,9)

plotter<-function(i){
temp.i<-fr[fr$per <=(i*.10),]
with(temp.i, plot(x, y, main=(i*.10),))
mod<-lm(y~x-1,data=temp.i)
r2<-summary(with(temp.i, lm(y~x)))$adj.r.squared
legend("bottomright", legend=signif(r2), col="black")
test[i]<-r2
print(r2)
abline(mod)
rm(temp.i)
rm(i)
rm(mod)
rm(r2)
}

test

Test comes up as the original vector of zeros. I know r2 is created for a
couple reasons (print works, and they show up on the graphs). Also, if I
run the function line by line with i assigned a number, test[i] is assigned
as it should be. However, if I call the function, plotter(i), test is not
modified, although the r^2 prints.

Mystified. What am I missing?

	[[alternative HTML version deleted]]

Hello,


btc1 wrote
> 
> Hello,
> 
> I'm trying to create a vector of r^2 values for using a function which
I
> will run in a "for" loop. Example:
> 
> per<-rnorm(100,.5,.2)^2
> x<-rnorm(100,10,5)
> y<-rnorm(100,20,5)
> fr<-data.frame(x,y,per)
> 
> test<-rep(0,9)
> 
> plotter<-function(i){
> temp.i<-fr[fr$per <=(i*.10),]
> with(temp.i, plot(x, y, main=(i*.10),))
> mod<-lm(y~x-1,data=temp.i)
> r2<-summary(with(temp.i, lm(y~x)))$adj.r.squared
> legend("bottomright", legend=signif(r2), col="black")
> test[i]<-r2
> print(r2)
> abline(mod)
> rm(temp.i)
> rm(i)
> rm(mod)
> rm(r2)
> }
> 
> test
> 
> Test comes up as the original vector of zeros. I know r2 is created for a
> couple reasons (print works, and they show up on the graphs). Also, if I
> run the function line by line with i assigned a number, test[i] is
> assigned
> as it should be. However, if I call the function, plotter(i), test is not
> modified, although the r^2 prints.
> 
> Mystified. What am I missing?
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help@ 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.
> 

You're changing a copy in the function's environment only, not in the
environment where it was created.
To have the changed copy, have the function return it.

(By the way, when the function returns, it rm 'temp.i', etc.)

There is a way of making global assignments but don't do this

test[i] <<- r2

without readind P. Burns, The R Inferno, Circle 6.

http://www.burns-stat.com/pages/Tutor/R_inferno.pdf


Hope this helps,

Rui Barradas


--
View this message in context:
http://r.789695.n4.nabble.com/scalar-assignment-within-a-vector-within-function-tp4517310p4517387.html
Sent from the R help mailing list archive at Nabble.com.

On Thu, Mar 29, 2012 at 3:49 PM, Benjamin Caldwell
<btcaldwell at berkeley.edu> wrote:
> Hello,
>
> I'm trying to create a vector of r^2 values for using a function which
I
> will run in a "for" loop. Example:
>
> per<-rnorm(100,.5,.2)^2
> x<-rnorm(100,10,5)
> y<-rnorm(100,20,5)
> fr<-data.frame(x,y,per)
>
> test<-rep(0,9)
>
> plotter<-function(i){
> temp.i<-fr[fr$per <=(i*.10),]
> with(temp.i, plot(x, y, main=(i*.10),))
> mod<-lm(y~x-1,data=temp.i)
> r2<-summary(with(temp.i, lm(y~x)))$adj.r.squared
> legend("bottomright", legend=signif(r2), col="black")
> test[i]<-r2
> print(r2)
> abline(mod)
> rm(temp.i)
> rm(i)
> rm(mod)
> rm(r2)
> }
>
> test
>
> Test comes up as the original vector of zeros. I know r2 is created for a
> couple reasons (print works, and they show up on the graphs). Also, if I
> run the function line by line with i assigned a number, test[i] is assigned
> as it should be. However, if I call the function, plotter(i), test is not
> modified, although the r^2 prints.
>
> Mystified. What am I missing?

Add the line

test

to the end of your function; this will cause the function to return
the value of the vector test.

Then call the function as

test = plotter(1)

or something.

You're missing the fact that in R all function arguments are passed by
value (think of them as being copied) and your function assigns a
value in a local copy of the vector test. This copy is discarded when
the function exits. The global copy is not modified. Of course, when
you step through the lines individually (not within a function call),
you work with the global test.

HTH

Peter

Ok, I've modified it as suggested so as to not pass to the global. Thanks
for the suggestions, and the link to the interesting reading.

On Fri, Mar 30, 2012 at 12:26 PM, William Dunlap <wdunlap@tibco.com>
wrote:
> I think you will be happier in the long run if you use the approach
> Peter suggested.  Do not use <<-, remove the calls to rm()
> from your function (because they do nothing useful -- all variables
> in a function disappear when the function is done), have it return
> its local variable 'test' and call it as
>   test5 <-  plotter(5)
>
> If you treat all variables as globals you will run into trouble.
>
> Bill Dunlap
> Spotfire, TIBCO Software
> wdunlap tibco.com
>
>
> > -----Original Message-----
> > From: r-help-bounces@r-project.org
[mailto:r-help-bounces@r-project.org]
> On Behalf
> > Of Benjamin Caldwell
> > Sent: Friday, March 30, 2012 12:09 PM
> > To: Peter Langfelder
> > Cc: r-help
> > Subject: Re: [R] scalar assignment within a vector within function
> >
> > Thanks all  - ended up going with
> > test<<-test
> >
> > On Thu, Mar 29, 2012 at 4:40 PM, Peter Langfelder <
> > peter.langfelder@gmail.com> wrote:
> >
> > > On Thu, Mar 29, 2012 at 3:49 PM, Benjamin Caldwell
> > > <btcaldwell@berkeley.edu> wrote:
> > > > Hello,
> > > >
> > > > I'm trying to create a vector of r^2 values for using a
function
> which I
> > > > will run in a "for" loop. Example:
> > > >
> > > > per<-rnorm(100,.5,.2)^2
> > > > x<-rnorm(100,10,5)
> > > > y<-rnorm(100,20,5)
> > > > fr<-data.frame(x,y,per)
> > > >
> > > > test<-rep(0,9)
> > > >
> > > > plotter<-function(i){
> > > > temp.i<-fr[fr$per <=(i*.10),]
> > > > with(temp.i, plot(x, y, main=(i*.10),))
> > > > mod<-lm(y~x-1,data=temp.i)
> > > > r2<-summary(with(temp.i, lm(y~x)))$adj.r.squared
> > > > legend("bottomright", legend=signif(r2),
col="black")
> > > > test[i]<-r2
> > > > print(r2)
> > > > abline(mod)
> > > > rm(temp.i)
> > > > rm(i)
> > > > rm(mod)
> > > > rm(r2)
> > > > }
> > > >
> > > > test
> > > >
> > > > Test comes up as the original vector of zeros. I know r2 is
created
> for a
> > > > couple reasons (print works, and they show up on the
graphs). Also,
> if I
> > > > run the function line by line with i assigned a number,
test[i] is
> > > assigned
> > > > as it should be. However, if I call the function,
plotter(i), test
> is not
> > > > modified, although the r^2 prints.
> > > >
> > > > Mystified. What am I missing?
> > >
> > >
> > > Add the line
> > >
> > > test
> > >
> > > to the end of your function; this will cause the function to
return
> > > the value of the vector test.
> > >
> > > Then call the function as
> > >
> > > test = plotter(1)
> > >
> > > or something.
> > >
> > > You're missing the fact that in R all function arguments are
passed by
> > > value (think of them as being copied) and your function assigns a
> > > value in a local copy of the vector test. This copy is discarded
when
> > > the function exits. The global copy is not modified. Of course,
when
> > > you step through the lines individually (not within a function
call),
> > > you work with the global test.
> > >
> > > HTH
> > >
> > > Peter
> > >
> >
> >       [[alternative HTML version deleted]]
> >
> > ______________________________________________
> > R-help@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.
>
	[[alternative HTML version deleted]]