CODEWITHSUNDEEP
rlistdataframe
user976991
user976991

Reputation: 411

Extracting a dataframe with unique elements from a list

Hi I got a list like that 

$`2`
  chr.pos  nt.pos CNV GRP
      1  783605   1   2
      1  888149   1   2
      1  991311   1   2
      1 1089305   1   2
      1 1177669   1   2

$`4`
 chr.pos  nt.pos CNV GRP
      2 1670488   1   4
      2 1758800   1   4

$`6`
 chr.pos  nt.pos CNV GRP
      2 1902924   1   6
      2 1978088   1   6

and I want to extract for each element, the unique chromosome, the CNV and the group and the highest and lowest nt.pos, so the output would be, I prefer a data frame

chr.pos  Start     End       GRP
 1         783605    1177669   2
 2         1670488   175880    4
 2         1902924   1978088   6

I tried with this 

results<-lapply(mylist, function(x){

return(as.data.frame(unique(x$chr.pos),range(x$nt.pos)[1],range(x$nt.pos)  [2],unique(x$GRP)))

}
)

But of course, what I got is a list.

Could you help me please?

Upvotes: 0

Views: 180

Answers (3)

Sven Hohenstein
Sven Hohenstein

Reputation: 81693

This does the trick: (Assuming dat is your list of data frames.)

structure(
    as.data.frame(cbind(do.call(rbind,
                                lapply(dat,
                                       function(x) c(x[["chr.pos"]][1], 
                                                     range(x[["nt.pos"]])))),
                        as.numeric(names(dat)))),
    .Names = c("chr.pos", "Start", "End", "GRP"))

#   chr.pos   Start     End GRP
# 2       1  783605 1177669   2
# 4       2 1670488 1758800   4
# 6       2 1902924 1978088   6

Upvotes: 1

A5C1D2H2I1M1N2O1R2T1
A5C1D2H2I1M1N2O1R2T1

Reputation: 193527

Assuming that your list is named "dat", as below:

dat <- read.table(header = TRUE, text = "chr.pos  nt.pos CNV GRP
1  783605   1   2
1  888149   1   2
1  991311   1   2
1 1089305   1   2
1 1177669   1   2
2 1670488   1   4
2 1758800   1   4
2 1902924   1   6
2 1978088   1   6")
dat <- split(dat, dat$GRP)

First, a question: do you really need it as a list, or can it just be a long data.frame? If it has to remain a list, perhaps try the following:

  1. sapply()

    data.frame(t(sapply(dat, function(x) 
  data.frame(chr.pos = unique(x["chr.pos"]), 
             Start = min(x["nt.pos"]), 
             End = max(x["nt.pos"]), 
             GRP = unique(x["GRP"])))))

  2. lapply()

    do.call(rbind, lapply(dat, function(x) 
  data.frame(chr.pos = unique(x["chr.pos"]), 
             Start = min(x["nt.pos"]), 
             End = max(x["nt.pos"]), 
             GRP = unique(x["GRP"]))))

Both will result in:

#   chr.pos   Start     End GRP
# 2       1  783605 1177669   2
# 4       2 1670488 1758800   4
# 6       2 1902924 1978088   6

Second, if it can be a long data.frame, then explore data.table():

library(data.table)
DaT <- data.table(do.call(rbind, dat), key = "GRP")
DaT[, list(chr.pos = unique(chr.pos),
           Start = min(nt.pos),
           End = max(nt.pos)), by = key(DaT)]
#    GRP chr.pos   Start     End
# 1:   2       1  783605 1177669
# 2:   4       2 1670488 1758800
# 3:   6       2 1902924 1978088

Upvotes: 3

user976991
user976991

Reputation: 411

Thanks Sven,

I did it in a similar way using this

N <- length(results)  #
DF <- data.frame(chr=rep(NA, N), Start=rep(NA, N), End=rep(NA,N), Group=rep(NA,N), stringsAsFactors=FALSE)

for (i in 1:length(results)){

  DF[i, ] <- c(unique(results[[i]]$chr.pos), range(results[[i]]$nt.pos)[1], range(results[[i]]$nt.pos)[2],unique(results[[i]]$GRP))

}

Upvotes: 0

Related Questions