R: selecting element from list

Question

I have two elements:

id1 <- "dog"
id2 <- "cat"

I want to extract any combination of these elements (dogcat or catddog) from a vector

L <- c("gdoaaa","gdobbb","gfoaaa","ghobbb","catdog")
L

I tried:

L[grep(paste(id1,id2,sep="")),L]
L[grep(paste(id2,id1,sep="")),L]

but this gives an error.

I would be grateful for your help in correcting the above.

`L[grep(paste(id1,id2,sep=""),L)] L[grep(paste(id2,id1,sep=""),L)]` — HubertL, Mar 08 '16 at 23:42
A simple non-regex solution could be `grepl(id1, L) & grepl(id2, L)`. You can add `fixed = TRUE` to both if efficiency is important. — David Arenburg, Mar 08 '16 at 23:44
I don't understand it, but apparently `grepl("(dog(cat)?)", L)` works courtesy of http://stackoverflow.com/questions/1177081/mulitple-words-in-any-order-using-regex — thelatemail, Mar 09 '16 at 00:32
That pattern matches `dogcat` but also `dog`, as in: `grepl("(dog(cat)?)", "dog")`. — effel, Mar 09 '16 at 01:09

effel · Accepted Answer · 2016-03-09T01:28:26.770

The error is from misplaced parentheses, so these minor variations on your code will work.

L[grep(paste(id1,id2,sep=""), L)]
# character(0)
L[grep(paste(id2,id1,sep=""), L)]
# [1] "catdog"

Alternatively this is a regex one-liner:

L[grep(paste0(id2, id1, "|", id1, id2), L)]
# [1] "catdog"

That and some patterns in the comments will also match dogcatt. To avoid this you could use ^ and $ like so:

x <- c("dogcat", "foo", "catdog", "ddogcatt")
x[grep(paste0("^", id2, id1, "|", id1, id2, "$"), x)]
# [1] "dogcat" "catdog"

1 Answers1