如何使用R删除字母

Question

I would like to remove letters and symbols from a line using R,

temp <- " 20.9°C \n              a 07" 

everything from  to the end, and get only

20.9

I have used this expression, but something fails...

temp2 <- ( gsub("°C \n*?","", temp,  ignore.case = FALSE, perl = TRUE))

Answer 1

Since you state you only want 20.9 from your string and everything after removed, this will do the trick.

> temp <- ' 20.9°C \n              a 07'
> temp2 <- sub('^ (.*?)Â[^Â]+', '\\1', temp)
> temp2
## [1] "20.9"

If you end up needing to extract temperatures, you could start off with something like this:

> library(gsubfn)
> temp <- ' -0.5°C \n \n   20.9°C      a +25.0°C'
> temp2 <- strapply(temp, '([+-]?[0-9]+\\.[0-9]+)')[[1]]
## [1] "-0.5"  "20.9"  "+25.0"

Based off your comment, how can I transform 25/08/14 to 25-08-2014 ?

gsub('([0-9]{2})/([0-9]{2})/([0-9]{2})', '\\1-\\2-20\\3', '25/08/14')
## [1] "25-08-2014"