R help - Sep 2007 - Q: appending to non-existent vector?

This is a real newbie question. What makes it worse is that I know
I've seen the answer somewhere, but I can no longer find it.

If I have a loop that is supposed to generate a vector piecemeal,
adding an element each time through the loop, what do I do to stop it
failing the first time around the loop, when the vector doesn't yet
exist (so I can't use the append() function)?

Something like this:

myvec <- NULL

while( condition ) {
  myvec <- c(myvec, additional stuff)
}

However, if you know ahead of time how long the vector will be (you are
adding 1 element at a time), then it is best to initialize the vector to
the correct length:

myvec <- numeric(1000)

for (i in 1:1000) {
	myvec[i] <- additional stuff
}

In the second case you create a vector of length 1000 then insert
numbers into it.  In the first case you first create a vector of length
1, then next time through you create a new vector of length 2, copy a
value into position 1 then insert the new value into position 2 then
give it the same name as the previous vector (allowing the previous
version to be garbage collected at some point), on the 3rd iteration you
create a new vector of length 3, copy 2 values and insert 1, etc.  You
can see that that can fragment memory and take unneeded time which is
why the second method is prefered.  The only time to use the first
method is if you don't know how long each piece of 'additional
stuff' is
and you know that you will only be doing the loop a few times.


Hope this helps,


-- 
Gregory (Greg) L. Snow Ph.D.
Statistical Data Center
Intermountain Healthcare
greg.snow at intermountainmail.org
(801) 408-8111
 
 
> -----Original Message-----
> From: r-help-bounces at r-project.org 
> [mailto:r-help-bounces at r-project.org] On Behalf Of D. R. Evans
> Sent: Friday, September 21, 2007 11:15 AM
> To: r-help at stat.math.ethz.ch
> Subject: [R] Q: appending to non-existent vector?
> 
> This is a real newbie question. What makes it worse is that I 
> know I've seen the answer somewhere, but I can no longer find it.
> 
> If I have a loop that is supposed to generate a vector 
> piecemeal, adding an element each time through the loop, what 
> do I do to stop it failing the first time around the loop, 
> when the vector doesn't yet exist (so I can't use the 
> append() function)?
> 
> ______________________________________________
> 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.
>

On 9/21/2007 1:15 PM, D. R. Evans wrote:
> This is a real newbie question. What makes it worse is that I know
> I've seen the answer somewhere, but I can no longer find it.
> 
> If I have a loop that is supposed to generate a vector piecemeal,
> adding an element each time through the loop, what do I do to stop it
> failing the first time around the loop, when the vector doesn't yet
> exist (so I can't use the append() function)?
You can create an empty vector to start.  Exactly how depends on what 
you're putting in it, but something like numeric(0) or list() should do 
what you want.

Of course, this is a very slow way to build results:  your code will run 
much faster if you allocate all the space you need at the beginning, and 
just fill in the values as you go, e.g.

x <- numeric(100)
i <- 0
while (more stuff) {
  i <- i+1
  x[i] <- stuff
}

Duncan Murdoch

On 21-Sep-07 17:15:08, D. R. Evans wrote:
> This is a real newbie question. What makes it worse is that I know
> I've seen the answer somewhere, but I can no longer find it.
> 
> If I have a loop that is supposed to generate a vector piecemeal,
> adding an element each time through the loop, what do I do to stop it
> failing the first time around the loop, when the vector doesn't yet
> exist (so I can't use the append() function)?
I always use NULL, as in:

x<-NULL
for(...){ ... ; x<-c(x,...)}

It has not let me down yet!
Best wishes,
Ted.

--------------------------------------------------------------------
E-Mail: (Ted Harding) <Ted.Harding at manchester.ac.uk>
Fax-to-email: +44 (0)870 094 0861
Date: 21-Sep-07                                       Time: 22:33:26
------------------------------ XFMail ------------------------------