Skip to content

Rearrange Data Frame

3 messages · Stella Xu, Sarah Goslee, Jim Lemon

Original
Stella Xu
# 
My question is about how to select and rearrange the data to a new data
frame
Here is an example:
Samples  counts  time 
A                 10           3
A                 12           4
A                 11           3
B                 12           4
B                 10           5
C                 11           2
C                 13           3
Say, if I want to make a new table that only look at ?counts? as
below:
A        B       C
10     12     11
12     10     13
11
How can I do this in R?
Thank you!
.
Sarah Goslee
# 
Hi Stella,
On Wed, Jul 29, 2015 at 1:14 PM, Stella Xu <Stella.Xu at hli.ubc.ca> wrote:
Your example data doesn't use time at all, and contains a duplicate
pair of A,?,3 - what do you want to have happen there? How should
duplicates be handled? How should the ordering of values work? If
instead that should be a 5, here's something that is almost what you
want (but I find more useful):

x <- structure(list(Samples = c("A", "A", "A", "B", "B", "C", "C"),
    counts = c(10L, 12L, 11L, 12L, 10L, 11L, 13L), time = c(3L,
    4L, 5L, 4L, 5L, 2L, 3L)), .Names = c("Samples", "counts",
"time"), class = "data.frame", row.names = c(NA, -7L))

library(reshape2)
dcast(x, time ~ Samples, value.var="counts", sum)
  time  A  B  C
1    2  0  0 11
2    3 10  0 13
3    4 12 12  0
4    5 11 10  0

If you want the results "scooted up", I think there was recently a
discussion on this list on doing so.

Sarah
Jim Lemon
# 
Hi Stella,
I think Sarah is correct in asking if that is what you really want,
but you can get a list similar to what you asked for like this:

sample_names<-unique(sx.df$Samples)
sx.lst<-list()
for(sn in 1:length(sample_names))
 sx.lst[[sn]]<-sx.df$counts[sx.df$Samples==sample_names[sn]]
names(sx.lst)<-sample_names

If you really want a data frame, you will have to coerce the number of
values in each element of the list to the same length:

maxlen<-max(unlist(lapply(sx.lst,length)))
for(i in 1:length(sx.lst)) {
 to_fill<-maxlen-length(sx.lst[[i]])
 if(to_fill > 0) sx.lst[[i]]<-c(sx.lst[[i]],rep(NA,to_fill))
}
sx2.df<-as.data.frame(sx.lst)

Jim
On Thu, Jul 30, 2015 at 6:48 AM, Sarah Goslee <sarah.goslee at gmail.com> wrote: