Spliting the character into parts

问题

I observe the following character:

  l <- "mod, range1 = seq(-m, n, 0.1), range2 = seq(-2, 2, 0.1), range3 = seq(-2, 2, 0.1)"

Using regular expressions in R I desire to split l in the following structure:

[1] "mod"                      "range1 = seq(-m, n, 0.1)"
[3] "range2 = seq(-2, 2, 0.1)" "range3 = seq(-2, 2, 0.1)"

Unfortunetely, I didn't find a proper way to overcome the problem, yet. Anyone has an idea how is it possible to acquire such an elegeant split?

回答1:

I really doubt you can do it with regular expression. You are trying to parse your string and so you need a parser, which is generally more powerful than a regex. I don't think it's general enough, but you can take advantage of the R parser and the alist class. Try:

res<-eval(parse(text=paste0("alist(",l,")")))
paste0(names(res),ifelse(names(res)!="","=",""),as.character(res))
#[1] "mod"                    "range1=seq(-m, n, 0.1)" "range2=seq(-2, 2, 0.1)"
#[4] "range3=seq(-2, 2, 0.1)"

Keep in mind that the regex proposed solutions fail if there are nested brackets. Try them and mine with:

l<-"mod, range1 = seq(-m, n, 0.1), range2 = seq(-2, exp(2), 0.1), range3 = seq(-2, 2, 0.1)"

to understand what I mean.

回答2:

Based on this regex, you can use str_extract_all from stringr,

library(stringr)
str_extract_all(l, '(?:[^,(]|\\([^)]*\\))+')
#[[1]]
#[1] "mod" " range1 = seq(-m, n, 0.1)" " range2 = seq(-2, 2, 0.1)" " range3 = seq(-2, 2, 0.1)"

or

trimws(unlist(str_extract_all(l, '(?:[^,(]|\\([^)]*\\))+')))
#[1] "mod" "range1 = seq(-m, n, 0.1)" "range2 = seq(-2, 2, 0.1)" "range3 = seq(-2, 2, 0.1)"

回答3:

Here is a base R option based on the pattern showed in the OP's post. Here we match all the characters starting from ( to the ), skip it and split by , followed by space.

strsplit(l, "\\([^)]+\\)(*SKIP)(*F)|, ", perl = TRUE)[[1]]
#[1] "mod"                      "range1 = seq(-m, n, 0.1)"
#[3] "range2 = seq(-2, 2, 0.1)" "range3 = seq(-2, 2, 0.1)"

Update

Using @nicola's 'l'

strsplit(l, ", (?=[[:alnum:]]+\\s+\\=)", perl = TRUE)[[1]]
#[1] "mod"                           "range1 = seq(-m, n, 0.1)"   
#[3]  "range2 = seq(-2, exp(2), 0.1)" "range3 = seq(-2, 2, 0.1)" 

and the previous 'l'

strsplit(l, ", (?=[[:alnum:]]+\\s+\\=)", perl = TRUE)[[1]]
#[1] "mod"                      "range1 = seq(-m, n, 0.1)" 
#[3] "range2 = seq(-2, 2, 0.1)" "range3 = seq(-2, 2, 0.1)"

来源：https://stackoverflow.com/questions/37811450/spliting-the-character-into-parts