Mercurial > repos > rhpvorderman > shm_csr

annotate check_unique_id.r @ 4:c8f02bce10d0 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload commit 05326cc2233eb0fa8beaf4198bd7c1519f2a0bf9"
author rhpvorderman
date Mon, 15 Nov 2021 14:04:10 +0000
parents 64d74ba01a7c
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
1 args <- commandArgs(trailingOnly = TRUE) #first argument must be the summary file so it can grab the
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
2
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
3 current_file = args[1]
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
4
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
5 current = read.table(current_file, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="", check.names=F)
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
6
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
7 if(!("Sequence number" %in% names(current))){
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
8 stop("First argument doesn't contain the 'Sequence number' column")
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
9 }
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
10
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
11 tbl = table(current[,"Sequence ID"])
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
12 l_tbl = length(tbl)
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
13 check = any(tbl > 1)
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
14
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
15 #if(l_tbl != nrow(current)){ # non unique IDs?
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
16 if(check){
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
17 print("Sequence.ID is not unique for every sequence, adding sequence number to IDs")
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
18 for(i in 1:length(args)){
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
19 current_file = args[i]
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
20 print(paste("Appending 'Sequence number' column to 'Sequence ID' column in", current_file))
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
21 current = read.table(current_file, header=T, sep="\t", fill=T, stringsAsFactors=F, quote="", check.names=F)
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
22 current[,"Sequence ID"] = paste(current[,"Sequence ID"], current[,"Sequence number"], sep="_")
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
23 write.table(x = current, file = current_file, quote = F, sep = "\t", na = "", row.names = F, col.names = T)
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
24 }
64d74ba01a7c "planemo upload commit 78d1fae87dbcf490e49a9f99e7a06de7328e16d4"
rhpvorderman
parents:
diff changeset
25 }