[R] how to find maximum for multivariable data

[David Winsemius]

On Sep 12, 2009, at 11:36 AM, caroline choong wrote:

> Dear all,
> I have a data set as follows:
>
>         ID               cycle.number      cycle.result
> 1     2525                      1                   38
> 2     2525                      2                   38
> 3     2525                      3                   25
> 4     2525                      4                   25
> 5     2525                      5                   25
> 6     2525                      6                   25
> 7     2531                      1                   38
> 8     2531                      2                   38
> 9     2078                      1                   38
> 10    2078                      2                   38
>
> I want to find out the maximum cycle.number for each ID, and later  
> find the
> corresponding cycle.result for that cycle.
>
> I have already managed to pull out the maximum cycle by using a for  
> loop:
>
> max.cycle <- vector()
> patients <- (levels(factor(ID)))
> for (i in 1:length(patients)) {
>
> max.cycle[i] <- max(cycle.number[(ID %in% patients[i] )])
>
> }
>
> But i would like to know if there is a better or more elegant way of  
> pulling
> out the maximum cycle.number for each ID?

There is:

?tapply

 > patients <- read.table(textConnection("        ID                
cycle.number      cycle.result
+ 1     2525                      1                   38
+ 2     2525                      2                   38
+ 3     2525                      3                   25
+ 4     2525                      4                   25
+ 5     2525                      5                   25
+ 6     2525                      6                   25
+ 7     2531                      1                   38
+ 8     2531                      2                   38
+ 9     2078                      1                   38
+ 10    2078                      2                   38"), header=TRUE)

 > tapply(patients$cycle.number, patients$ID, max)
2078 2525 2531
    2    6    2


> Perhaps without the need for using
> a for loop?
>
> Many thanks,


David Winsemius, MD
Heritage Laboratories
West Hartford, CT