How to split a decimal number from a string in R

Question

For example:

c("2.11abc","15.1cde",".562342rfg")

How to split the numeric out of this vector? The vector I would like to have is

c(2.11, 15.1, 0.562342)

I tried

gsub("[^[:digit:]]", "", a)

but the result was

c(211, 151 ,562342)

I really appreciate your help.

Answer 1

If have more troublesome strings, in which periods can appear outside of the number part of the string, you might want to use something like this.

library(gsubfn)

## A possibly more realistic character vector
x <- c("2.11abc.def","a.b-15.1cde",".562342rfg", "abcdef")

getNumberPart <- function(x) {
    pat <- "(-?(\\d*\\.*\\d+|\\d+\\.))"
    strapply(x, pattern=pat, FUN=as.numeric, simplify=TRUE, empty=NA)
}

getNumberPart(x)
# [1]  2.110000 -15.100000   0.562342         NA

Do note that if a string contains more than one number, strapply() will extract all of them, and will return the overall result as a list (rather than a simple vector) with one list element per input string.

Answer 2

You are missing the period:

as.numeric(gsub("[^[:digit:].]", "", c("2.11abc","15.1cde",".562342rfg")))
# [1]  2.110000 15.100000  0.562342