| To: | "s-news@utstat.toronto.edu" <s-news@utstat.toronto.edu> |
|---|---|
| Subject: | Summary(2): how to generate the sequence |
| From: | Ning Li <lin@cryptic.rch.unimelb.edu.au> |
| Date: | Wed, 28 Mar 2001 13:33:30 +1000 |
|
Dear lister, Many thanks to Thomas Jagger Phd from Insightful Corporation. He contributed one more method to create a sequential id. The three types of methods we have had so far are: 1) -rep(1:n, each=k)- where n=unique(id), k=# of repeated measures each id 2) -as.numeric(as.factor(id))- 3) -match(id,unique(id))- The argument -each- in -rep()- is usually a constant, meaning that method 1) works if all id were observed for the same number of times (no missing values in a balanced design). The 2) and 3) are more general in this sense. Best Regards ------------------------------------------------------------------- Ning Li Research Assistant, CEBU Univ. of Melbourn |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: No commands for barplots with error..., David M Smith |
|---|---|
| Next by Date: | S-Plus on Mac OS X?, P . J . Wells |
| Previous by Thread: | Re: No commands for barplots with error..., Sung, Myong-Hee (NCI) |
| Next by Thread: | S-Plus on Mac OS X?, P . J . Wells |
| Indexes: | [Date] [Thread] [Top] [All Lists] |