ansaurus

Question

Answer 1

+1 A:

you can easily accomplish this using the excellent reshape/ reshape2 package by hadley. here is the code to take you to what you need

library(reshape); 
df = melt(df, id = c('subj', 'assessment'));
df = cast(df, subj ~ variable + assessment);

let me know if this works for you.

Ramnath 2010-09-20 16:32:27

I believe that should be "variable" instead of "test" in the cast formula, like so: df = cast(df, subj ~ variable + assessment). (Unless you add variable_name = "test" to the previous line).

Fojtasek 2010-09-20 16:48:35

that is right. thanks for pointing it out. i have modified the code accordingly

Ramnath 2010-09-20 20:25:37

Answer 2

+1 A:

The reshape function (in stats) does this fairly easily:

reshape(data, timevar='assessment', idvar='subj', dir='wide')

Or to just get the results for test1:

reshape(subset(data, select=-test2), timevar='assessment', idvar='subj', dir='wide')

Charles 2010-09-20 16:38:40

Answer 3

A:

Thanks! This is why OpenSource is so great. The first method (reshape{stats}) worked best for me once I figured out how to use "v.names" and "varying" (otherwise, you get a redundant all-to-all reshaping,- bad in a dataset with 200 variables).

THanks again

JOn

Jon Erik Ween 2010-09-22 15:25:07

ansaurus

tags:

views:

answers:

R reorder dataframe within subject

related questions