On 01/22/2013 07:19 PM, Troels Ring wrote:> Dear friends - this is a very simple question - I have a data frame
> 'data.frame': 87 obs. of 3 variables:
> $ ID : int 1 1 1 2 2 2 3 3 3 4 ...
> $ prep : num 1.18 1.38 1.34 1.93 2.38 2.24 1.17 1.13 1.21 1.89 ...
> $ postp: num 0.63 0.71 0.75 1.01 1.12 1.07 0.87 0.64 0.7 0.8 ...
>
> - 29 persons (ID) each measured three times before and after an
> intervention: prep and postp -
> I need data rearranged like
>
> ID time val
> 1 1 prep
> 1 2 postp
> 1 1
> 1 2
> 1 1
> 1 2
> I cannot make reshape or stack do the trick.
>
Hi Troels,
With a bit of extra processing I think rep_n_stack (prettyR) will do
what you want:
# fake some data
tr.df<-data.frame(ID=rep(1:29,each=3),prep=runif(87,1,3),postp=runif(87,0.5,1.5))
# add a repeat number
tr.df$repno<-rep(1:3,29)
# get the reshaped data frame
trlong.df<-rep_n_stack(tr.df,to.stack=2:3,
stack.names=c("prepost","value"))
# reorder it
trlong.df[order(trlong.df$ID,trlong.df$repno),]
ID repno prepost value
1 1 1 prep 2.9158693
88 1 1 postp 0.9932342
2 1 2 prep 1.2852817
89 1 2 postp 0.8187234
3 1 3 prep 2.5771902
90 1 3 postp 1.0033936
4 2 1 prep 2.2969320
91 2 1 postp 0.6837140
5 2 2 prep 1.3083553
92 2 2 postp 1.4537096
6 2 3 prep 2.8654184
93 2 3 postp 1.0880881
...
Jim