Is the result of FindStringSubmatch correct?

[sleepy]

https://play.golang.org/p/YoNyiKo-lTc

package main

import (

"fmt"

"regexp"

)

func main() {

s := `([^ ]+ ?)*`

re := regexp.MustCompile(s)

fmt.Printf("%q\n", re.FindStringSubmatch("aaa bbb ccc"))

}

I expect the output to be:

["aaa bbb ccc" "aaa " "bbb " "ccc"]

but the actual output is:

["aaa bbb ccc" "ccc"]

 

 

[Brian Candler]

Perhaps you were confused by this in the documentation:

"...and the matches, if any, of its subexpressions, as defined by the 'Submatch' descriptions in the package comment"

That means one value per actual parenthesised subexpression in the original expression. e.g. given expression ((\S+) (\S+) (\S+)) the results will be "aaa bbb ccc", "aaa", "bbb", "ccc".

((\S+) (\S+) (\S+)) 

^^     ^     ^

||     |     |

|$2    $3    $4
|

$1

But if one parenthesised expression matches multiple times due to a repeat, it still gives a single value (the last match).  That is: $1 is always $1, and $2 is always $2, based on their positions in the regexp.  They don't shift along if there are multiple matches.

Maybe you want FindAllSubmatch instead.

https://play.golang.org/p/vOIvg-r4EI1

[Brian Candler]

Or closer to your original code, FindAllStringSubmatch.

https://play.golang.org/p/Tiva8X425k3

[bomin]

Got it, thanks.

在 2020年3月2日星期一 UTC+8下午7:39:28，Brian Candler写道：