[R] melting a data frame

[Data Analytics Corp.]

Hi,

I received several really good suggestions, but the feeling seems to be 
to use a for loop.  I did this at first, but my curiosity took over -- 
hence, my question.  The loop, coupled with a list of relevant columns, 
seems best.

Thanks all,

Walt

________________________

Walter R. Paczkowski, Ph.D.
Data Analytics Corp.
44 Hamilton Lane
Plainsboro, NJ 08536
________________________
(V) 609-936-8999
(F) 609-936-3733
walt at dataanalyticscorp.com
www.dataanalyticscorp.com
_____________________________________________________


On 9/8/2013 10:06 AM, John Fox wrote:
> Dear Walt and A.K.,
>
> One shouldn't reflexively avoid loops in R. In this case, it seems to me clearer to use a loop, and it's no less "efficient" (especially, I would guess, when one takes into account the time to figure out how to do the computation). I get
>
>> system.time({
> + res<-do.call(rbind,lapply(split(colnames(dat1),((seq_len(ncol(dat1))-1)%/%21)+1),function(x) {x1<- dat1[,x]; colnames(x1)<- paste("V",1:21);x1}))
> + row.names(res)<- 1:nrow(res)
> + })
>     user  system elapsed
>     0.02    0.00    0.02
>> dim(res)
> [1] 1170   21
>
>
>> system.time({
> + res2 <- as.data.frame(matrix(0, 1170, 21))
> + for (i in 1:9){
> +     res2[((i - 1)*130 + 1):(i*130), ] <- dat1[, ((i - 1)*21 + 1):(i*21)]
> + }
> + })
>     user  system elapsed
>     0.02    0.00    0.01
>> dim(res2)
> [1] 1170   21
>
>> all(res == res2)
> [1] TRUE
>
> Best,
>   John
>
> On Sun, 8 Sep 2013 06:29:42 -0700 (PDT)
>   arun <smartpink111 at yahoo.com> wrote:
>>
>> Hi,
>>
>> You could try:
>> set.seed(48)
>> dat1<- as.data.frame(matrix(sample(1:40,189*130,replace=TRUE),ncol=189))
>> res<-do.call(rbind,lapply(split(colnames(dat1),((seq_len(ncol(dat1))-1)%/%21)+1),function(x) {x1<- dat1[,x]; colnames(x1)<- paste("V",1:21);x1}))
>>   row.names(res)<- 1:nrow(res)
>>   dim(res)
>> #[1] 1170   21
>> A.K.
>>
>>
>>
>> ----- Original Message -----
>> From: Data Analytics Corp. <walt at dataanalyticscorp.com>
>> To: R help <r-help at r-project.org>
>> Cc:
>> Sent: Saturday, September 7, 2013 11:33 PM
>> Subject: [R] melting a data frame
>>
>> Hi,
>>
>> Suppose I have a data frame with 189 columns.  The columns are actually 9 blocks of 21 columns each, each block representing measures on each of 9 products.  There are 130 rows.  Suppose I extract the first block of 21 columns and make them into a separate data frame.  I then want to take the second block of 21 columns and rbind it to the first; then the third set of 21 and rbind it to the first two; etc.  The final data frame should have 1170 (= 9 * 130)  rows and 21 columns.  Is there an easy way to melt the blocks comparable to using the melt function in the plyr package (which is why I'm referring to what I want to do as "melting")?  It seems that there should be a simple way to do this.  I used a for loop which worked, but I want to see if there's a more efficient way.
>>
>> Thanks,
>>
>> Walt
>>
>> ________________________
>>
>> Walter R. Paczkowski, Ph.D.
>> Data Analytics Corp.
>> 44 Hamilton Lane
>> Plainsboro, NJ 08536
>> ________________________
>> (V) 609-936-8999
>> (F) 609-936-3733
>> walt at dataanalyticscorp.com
>> www.dataanalyticscorp.com
>>
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>>
>>
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
> ------------------------------------------------
> John Fox
> McMaster University
> Hamilton, Ontario, Canada
> http://socserv.mcmaster.ca/jfox/
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>