Skip to content

reshape a dataset

6 messages · Jim Lemon, Sohail Khan, Jorge I Velez

Sohail Khan
# 
Hi
I have data set as follows:
  A 92315  A 35018  A 56710  B 52700  B 92315  B 15135  C 35018  C 52700
I would like to transform this data set into:
  ID 92315 35018 56710 52700 15135  A 1 1 1 0 0  B 1 0 0 1 1  C 0 1 0 1 0
I looked into reshape package to no avail.
I would appreciate any suggestions.

-Sohail
Jim Lemon
#
On Thu, 14 Aug 2014 06:08:51 PM Sohail Khan wrote:
52700
1 0
Hi Sohail,
You are doing a bit more than reshaping. This may get you there:

skdat<-read.table(text="A 92315
A 35018
A 56710
B 52700
B 92315
B 15135
C 35018
C 52700",stringsAsFactors=FALSE)
names(skdat)<-c("lettertag","ID")
ID<-unique(skdat$ID)
lettertags<-unique(skdat$lettertag)
newskdat<-list(ID)
for(i in 1:length(lettertags))
 newskdat[[i+1]]<-
  as.numeric(ID %in% skdat$ID[skdat$lettertag==lettertags[i]])
names(newskdat)<-c("ID",lettertags)

I'm assuming that you don't really want your answer as a single string.

Jim
Jorge I Velez
# 
Dear Sohail,

Using Jim's data set skdat, two more options would be

# first option
d <- with(skdat, table(ID, lettertag))
names <- colnames(d)
d <- c(list(rownames(d)), lapply(1:ncol(d), function(i) as.numeric(d[,i])))
names(d) <- c('ID', names)
d

# second option
d <- with(skdat, table(ID, lettertag))
res <- c(list(rownames(d)), sapply(apply(d, 2, list), "[", 1))
names(res)[1] <- "ID"
res

HTH,
Jorge.-
On Fri, Aug 15, 2014 at 7:19 PM, Jim Lemon <jim at bitwrit.com.au> wrote:

            

      
      
      
    

        

      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Sohail Khan
    

    

      
      #
    

  


  

      
Thanks Jim and Jorge,
Clever solutions, the final output is a list.
How do I covert it back a dataframe?
-Sohail


On Fri, Aug 15, 2014 at 5:37 AM, Jorge I Velez <jorgeivanvelez at gmail.com>
wrote:

            

      
      
      
    

        

      
      
    

            

      
      
      
    

        

      
      
    

        

  
  


  


      

    

    

      
      

        


  

        

      Jim Lemon
    

    

      
      #
    

  


  

      
On Fri, 15 Aug 2014 06:22:53 AM Sohail Khan wrote:

            

      
      
      
    

        
as.data.frame(newskdat)

or for Jorge's

as.data.frame(res)

Jim

  
  


  


      

    

    

      
      

        


  

        

      Jorge I Velez
    

    

      
      #
    

  


  

      
If that's the case, you could do the following:

d <- with(skdat, table(ID, lettertag))
d <- data.frame(cbind(ID = rownames(d), d))
rownames(d) <- NULL
d

HTH,
Jorge.-

        
On Fri, Aug 15, 2014 at 8:22 PM, Sohail Khan <sohail13 at gmail.com> wrote: