What you need to do is just separate the dose records, merge with the original dataset before sorting it:
A few comments:
1. There may be other ways to do this task. I prefer this one because I can modify the separated dose records (change CMT, RATE, etc) for different absorption models.
2. I do not know what is your data source. In my experience, the dose and concentration data almost always come from different sources. If I need to do the data management myself, I will keep those 2 sources of data separated right before I finish. That gives you more flexibility.
3. ASCII files are preferred than xls as it is easy to be imported into R and can be used by NM.
Glad to see another NMuser.