On 5/18/2011 11:22 AM, Nick Holford wrote:
If you use R or SAS to skip the first TABLE and header records you still
need to figure out how to process the subsequent TABLE and header
records if you your table file has multiple tables as it often does when
it is created by simulation.
I keep forgetting how to do this efficiently... this is one way to read
nonmem simulation results (with ONEHEADER NOPRINT FILE=sdtab101):
res.101<-read.table("sdtab101",head=T,skip=1,fill=T,as.is=T)
res.101<-res.101[!is.na(as.numeric(res.101[,1])),]
The important arguments to read.table are "fill", which makes R accept
non-rectangular data, and "as.is", which prevents R from converting
character data to factors.
Later, res.101 should probably be converted to numeric, but that's a
different story...
(maybe somebody knows a more elegant solution to this?)
--
Paul Matthias Diderichsen, PhD
Quantitative Solutions B.V.
+31 624 330 706
