I found a solution:
pairs<-merge(subset(data.frame,cond==2,select=c(group,x),subset(data.frame,cond==3,select=c(group,x)),by=c("group"))
That yields something like
group cond.x cond.y
A 4 6.5
D 5 6
E 4 6
now I just need to figure out how I embed this into a function together
with the test, so that I don't have to write to much ....
Stefan Grosse
> Dear list,
>
> I want to extract pairs of values out of a dataframe where one
> criteria/condition does match.
>
> I have an experiment with 3 conditions which were not always applied:
>
> e.g.:
>
> group cond x
> A 1 2
> A 2 4
> A 3 6.5
> B 1 3
> B 2 4.5
> C 1 2.5
> C 3 4
> D 2 5
> D 3 6
> E 1 1
> E 2 4
> E 3 6
>
>
> Now I wanted to extract the x of those groups where condition 2 and
> condition 3 do both exist.
>
> In this example that would be groups A, D and E and the extracted pairs
> e.g.:
> cond2 cond3
> 4 6.5
> 5 6
> 4 6
>
>