Use strsplit with multiple delimiters

Question

How can I split this

 Chr3:153922357-153944632(-)
 Chr11:70010183-70015411(-)

in to

    Chr3  153922357 153944632 - 
    Chr11 70010183  70015411  -

I tried strsplit(df$V1,"[[:punct:]]")), but the negative sign is not coming in the final result

(I think) I accidentally deleted a valid post below (not mine)! My deepest apologies! I can't remember who's it was. I flagged this for the mod's attention to undelete. — Maurits Evers, Dec 12 '17 at 13:33
You want to split only on the first occurrence of `-`. Perhaps some of the answers [here](https://stackoverflow.com/questions/26246095/r-strsplit-on-first-instance) may help — KenHBS, Dec 12 '17 at 13:33

Alex Gold · Answer 1 · 2017-12-12T13:49:50.967

2

The issue is that - is both a character you want to extract and a delimiter. Your best bet is using capture groups and specifying the full regex string:

stringr::str_match(x, "^(.{4}):(\\d+)-(\\d+)\$(.)\$$")

EDIT: If you want to let the first capture group capture strings of arbitrary length (e.g. ChrX for any X), you can change first capture group from .{4} to Chr\\d+.

edited Dec 12 '17 at 13:49

answered Dec 12 '17 at 13:28

Alex Gold

335
1
9

if i use stringr::str_match(x, "^(.{4}):(\\d+)-(\\d+)\$(.)\$$"), the second row is getting NA – Kryo Dec 12 '17 at 13:34
Regex should cover cases for ChrNNA, i.e.: chr1, chr12, chrX. – zx8754 Dec 12 '17 at 13:41

Maurits Evers · Answer 2 · 2017-12-12T13:43:30.340

How about this in base R using stringsplit and gsub:

# Your sample strings
ss <- c("Chr3:153922357-153944632(-)",
        "Chr11:70010183-70015411(-)")

# Split items as list of vectors 
lst <- lapply(ss, function(x)
    unlist(strsplit(gsub("(.+):(\\d+)-(\\d+)\\((.)\\)", "\\1,\\2,\\3,\\4", x), ",")))


# rbind to dataframe if necessary
do.call(rbind, lst);
#    [,1]    [,2]        [,3]        [,4]
#[1,] "Chr3"  "153922357" "153944632" "-"
#[2,] "Chr11" "70010183"  "70015411"  "-"

This should work for other chromosome names and positive strand features as well.

score 2 · Accepted Answer · answered Dec 12 '17 at 14:08

You can also try str_split from stringr:

library(stringr)
lapply(str_split(df$V1, "(?<!\\()\\-|[:\\)\\(]"), function(x) x[x != ""])

Result:

[[1]]
[1] "Chr3"      "153922357" "153944632" "-"        

[[2]]
[1] "Chr11"    "70010183" "70015411" "-"

Data:

df = read.table(text = " Chr3:153922357-153944632(-)
 Chr11:70010183-70015411(-) ")

Use strsplit with multiple delimiters

3 Answers3