[R] A problem about outer()

[Feng Chen]

Thanks for the informative comments and the detailed explanation.

----- Original Message ----- 
From: "Adaikalavan Ramasamy" <ramasamy at cancer.org.uk>
To: "Feng Chen" <fengchen at hkusua.hku.hk>
Cc: "R-help" <r-help at stat.math.ethz.ch>
Sent: Tuesday, March 01, 2005 12:25 AM
Subject: Re: [R] A problem about outer()


> You might want to read (or re-read) the posting guide about giving a
> simple example. See comments below.

Henceafter, I won't post any question before:
1. searching the documentation,
2. googling the web,
3. reading the code of the relevant functions,
and
4. a careful rereading of the posting guide.

>
>
> On Mon, 2005-02-28 at 23:03 +0800, Feng Chen wrote:
>> Dear all,
>>
>> I have something about function outer() that I can't understand. Just 
>> see the following example. The two NaNs are due to 0/0, but I can't 
>> figure out the cause of the last two errors. I wonder if some one can 
>> explain this for me.
>> ___________________________________________________________________
>> > sx=rbinom(10,1,0.5);ot=rbinom(10,1,0.5);ag <- rbinom(10,100,0.3);ho 
>> > <- rbinom(10,100,0.5)
>
>
> Cute but unfortunately not very legible. Why are you mixing "=" and
> "<-" ? Is there a problem with space bars and return key on your

I am not sure about the reason for that mixture. Perhaps because "=" and 
"_" are so close on the keyboard. Anyway, they were definitely not made 
so by intention.

> keyboard ? Please learn to wrap the emails at about 72 characters per
> line (see http://expita.com/nomime.html).

Thanks, I have followed the instructions listed on that page.
>
>
>> > dp <- 
>> > function(s,a,h)sum((sx==s)&(ag==a)&(ho==h)&(ot==1))/sum((sx==s)&(ag==a)&(ho==h))
>
>> > (function(x,y)dp(1,x,y))(2,3)
>> [1] NaN
>> > (function(x,y)dp(0,x,y))(27,52)
>> [1] NaN
>
> Again this is confusing. Why not define another function (you will 
> need

Here just for testing pupose. I thought I just need that once (in the 
outer() function).

> to anyway - see below).
>
> Alternatively, you can set 1 as the default value for s in dp().
>
>> > dpm <- outer(ag,ho,function(x,y)dp(1,x,y))
>> Error in outer(ag, ho, function(x, y) dp(1, x, y)) :
>>  dim<- : dims [product 100] do not match the length of object [1]
>
>>From help("outer") :
>
> 'FUN' must be a function (or the name of it) which expects at
> least two arguments and which operates elementwise on arrays.
>
>
> And following the suggestion of Prof. Daalgard in the thread
> http://tolstoy.newcastle.edu.au/R/help/00a/1445.html
>
> dp.vect <- function(s, x, y){
>  sapply( 1:length(x), function(i) dp( s=s, a=x[i], h=y[i]) )
> }
> outer(ag, ho, FUN=dp.vect, s=1 )
> # works but I leave the verification to you

It does work.

>
>
>
> Your problem could be generalised as the following example
>
> one <- rnorm(3); two <- rnorm(4)                 # data
> outer( one, two, function(x, y) x + y )          # works fine
> outer( one, two, function(x, y) sum(c(x, y)) )   # does not work

I had tried this before posting.

>
> sum.vect <- function(x, y){
>  sapply( 1:length(x), function(i) sum( c( x[i], y[i] ) ) )
> }
> outer( one, two, sum.vect )
>

You have thoroughly explained away my confusion. Thanks for your great 
patience.

Thank all for being so tolerant and friendly to newcomers.

>
>
>> > dpf <- outer(ag,ho,function(x,y)dp(0,x,y))
>> Error in outer(ag, ho, function(x, y) dp(0, x, y)) :
>>  dim<- : dims [product 100] do not match the length of object [1]
>> >
>> -------------------------------------------------------------------------------------------------------------------
>>
>> Thanks very much,
>> Feng
>>
>> [[alternative HTML version deleted]]
>>
>> ______________________________________________
>> R-help at stat.math.ethz.ch mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide! 
>> http://www.R-project.org/posting-guide.html
>>