How to Un-group a grouped data set?

-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-
project.org] On Behalf Of R. Michael Weylandt
Sent: Tuesday, May 15, 2012 1:09 AM
To: Cheenghee AM Koh
Cc: r-help at r-project.org
Subject: Re: [R] How to Un-group a grouped data set?

It is a nifty and surprisingly useful construct whenever you need to
construct a function call programmatically or apply it to a list.

R-News 2/2 has some useful tips on this and related functions in the
Programmer's Note section if you're interested.

Best,
Michael

On Tue, May 15, 2012 at 2:05 AM, Cheenghee AM Koh <sigontw at gmail.com>
wrote:

Thank you so much! ?I can't believe I spent the whole night by not

knowing

this one command "do.call"
This is so handy!
Best, Koh


On Tue, May 15, 2012 at 12:52 AM, R. Michael Weylandt
<michael.weylandt at gmail.com> wrote:

Sorry -- I missed the bit about the AE in your original post.

Perhaps

you can work with my bit for the repeats, but it looks like if you
want to use your function, it should suffice to do something like

do.call("rbind", lapply(NewFuncName, 1:6))

Best,
Michael

On Tue, May 15, 2012 at 1:50 AM, R. Michael Weylandt
<michael.weylandt at gmail.com> wrote:

Don't use subset for a function name -- it's already the name of a
rather important function as is data (but at least that one's not

a

function in your use so it's not quite so bad). Finally, use

dput()

when sending data so we get a plaintext reproducible version.

I'd try something like this:

dats <- structure(list(Study = c(1L, 1L, 2L, 2L, 3L, 3L), TX =

c(1L,

0L, 1L, 0L, 1L, 0L), AEs = c(3L, 2L, 1L, 2L, 1L, 1L), N = c(5L,
7L, 10L, 7L, 8L, 4L)), .Names = c("Study", "TX", "AEs", "N"),

class =

"data.frame", row.names = c("1",
"2", "3", "4", "5", "6"))

# See how handy dput can be :-)

dats[unlist(mapply(FUN = function(x,y) rep(x, y), 1:NROW(dats),
dats$N)), -4]

which isn't super elegant, but others might have something better.

Best,
Michael

On Tue, May 15, 2012 at 1:24 AM, Cheenghee AM Koh

<sigontw at gmail.com>

wrote:

Hello, R-fellows,

I have a question that I really don't know how to solve. I have

spent

hours
on line surfing for possible solutions but in veil. Please if

anyone

could
help me handle this issue, you would be so appreciated!

I have a "grouped" dataset like this:

data

?Study TX AEs ? N
1 ? ? 1 ? ? 1 ? ?3 ? ? ? 5
2 ? ? 1 ? ? 0 ? ?2 ? ? ? 7
3 ? ? 2 ? ? 1 ? ?1 ? ? ?10
4 ? ? 2 ? ? 0 ? ?2 ? ? ? 7
5 ? ? 3 ? ? 1 ? ?1 ? ? ? 8
6 ? ? 3 ? ? 0 ? ?1 ? ? ? 4

where Study is the study id, TX is treatment, AEs is how many

people in

this trial is positive, and N is the number of the subjects.

Therefore,

for
the row 1, it stands for: It is the treatment arm for the study

one,

where
there are 5 subjects and 3 of them are positive. The row 2 stands

for:

It
is the control arm of the study 1 where there are 7 subjects and

2 of

them
are positive.

Now I would like to "un-group them", make it like:

Study ?TX ? AEs
? 1 ? ? ? ? 1 ? ? ?1
? 1 ? ? ? ? 1 ? ? ?1
? 1 ? ? ? ? 1 ? ? ?1
? 1 ? ? ? ? 1 ? ? ?0
? 1 ? ? ? ? 1 ? ? ?0
? 1 ? ? ? ? 0 ? ? ?1
? 1 ? ? ? ? 0 ? ? ?1
? 1 ? ? ? ? 0 ? ? ?0
? 1 ? ? ? ? 0 ? ? ?0
? 1 ? ? ? ? 0 ? ? ?0
? 1 ? ? ? ? 0 ? ? ?0
? 1 ? ? ? ? 0 ? ? ?0
? 2 ? ? ? ? 1 ? ? ?1
? .....................
?.....................


But I wasn't able to do it. In fact I wrote a small function, and

use

"lapply" to get what I want. It worked well, and did give me what

I

want.
But I wasn't able to collapse all the returns into one single

data

frame
for subsequent analysis.

The function I wrote:

subset = function(i){
d = c(rep(data[i,1], data[i,4]), rep(data[i,2], data[i,4]),

rep(0:1,

c(data[i,4] - data[i,3],data[i,3])))
d = matrix(d, data[i,4],3)
d
}

then:

Data = lapply(1:6, subset)
Data

Therefore, I tried to write a loop. But no matter how I tried, I

can't

get
what I want.

Any idea?

Thank you so much!

Best,


--
Cheenghee Masaki Koh, MSW, MS(c), PhD Student
School of Social Service Administration
Department of Health Studies, Division of Biological Science
University of Chicago

? ? ? ?[[alternative HTML version deleted]]

code.

--
Cheenghee Masaki Koh, MSW, MS(c), PhD Student
School of Social Service Administration
Department of Health Studies, Division of Biological Science
University of Chicago