Summary: how to generate the sequence

[Ning Li <lin@cryptic.rch.unimelb.edu.au>]

~ ~ ~ ~ ~ ~ ~ extract of the original query: ~ ~ ~ ~ ~ ~ ~ 

I have a variable id looks like this:
[1] 920006 920006 920006 920006 920006 920006
[7] 920009 920009 920009 920009 920009 920009
[13] 920020 920020 920020 920020 920020 920020

How to generate a corresponding variable, dumyid, as
1 1 1 1 1 1 2 2 2 2 2 2 3 3 3 3 3 3

 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 

Dear lister,

I would like to thank S.D.Byers, Charles E. Wright, Bill Venerables, John
Maindonald, Nick Ellis, 
Briand D. Ripley, Piotrovskij Vladimir, Bjarke Klein and many more, for
helping me in finding 
a solution to the above question.  

There are two types of solutions. One is to directly generate the
sequence using -rep()- :
                  
-sort(rep(c(1:n),k))-    or    -rep(1:n,
each=k)- 
where n=the number of unique id,  k=the number of repeated
measurement for each id.
 
An alternative is to use -factor- on the original id variable. This works
because factors are 
internally recorded as integers.
      -codes(factor(id))-   or 
-as.numeric(factor(id))-   or  
-as.numerical(as.factor(id))-

The command -unclass(factor(id))- also generates integer indexes of the
categories, but 
it returns the original attributes as well. So it doesn't quite fit in
this the case.  Especially 
the -seq()- command which I had been trying on doesn't work at all here.

Best Regards

-------------------------------------------------------------------
Ning Li
Research Assistant, CEBU  Univ. of Melbourne