How to split on a character only if it's outside of quotes in golang?

Question

I need to split a chunk of text on the + symbol, but only when it's outside of single quotes. The text will look something like:

Some.data:'some+value'+some.more.data:9+yet.more.data:'rock+roll'

which should become a slice of three values:

Some.data:'some+value'
some.more.data:9
yet.more.data:'rock+roll'

I've found similar questions that do it using regex, but that requires look ahead which the golang regex engine doesn't have.

I also took a crack at creating my own regex without lookahead:

'.*?'(\+)|[^']*(\+)

But that seems to fall apart on the third item where it splits on the + in 'rock+roll'.

I've thought about potentially doing a string split on + and then validating each slice to make sure it's not a partial expression and then stitching the pieces back together if it is, but it will be fairly involved and i'd like to avoid it if possible.

At the moment I think the best solution would be to identify text that is inside of quotes (which I can easily do with regex), either URL encode that text or do something else with the plus sign, split the text and then URL decode the expression to get the + sign inside of quotes back, but i'm wondering if there is a better way.

Does anyone know of a way to split on a + sign that is outside of quotes using regex without lookahead? Can anyone think of a simpler solution than my URL encoding/decoding method?

score 4 · Accepted Answer · answered Sep 28 '21 at 16:27

4

Plain code can be easier:

func split(s string) []string {
    var result []string
    inquote := false
    i := 0
    for j, c := range s {
        if c == '\'' {
            inquote = !inquote
        } else if c == '+' && !inquote {
            result = append(result, s[i:j])
            i = j +1
        }
    }
    return append(result, s[i:])
}

answered Sep 28 '21 at 16:27

Daddy likes.... – Abe Miessler Sep 28 '21 at 16:29
1

@AbeMiessler Be careful with escaped quotes. A value of `Some.data:'Couer d\\'Alene'` could prove problematic. – sberry Sep 28 '21 at 17:22
@gopher - I already updated the function in my local implementation to handle escaping, but thanks! – Abe Miessler Sep 28 '21 at 21:51

score 0 · Answer 2 · answered Jun 24 '23 at 10:55

Slightly old question, but you might consider using go-andiamo/splitter

Example:

package main

import (
    "github.com/go-andiamo/splitter"
)

func main() {
    plusSplitter := splitter.MustCreateSplitter('+', splitter.SingleQuotes)

    s := "Some.data:'some+value'+some.more.data:9+yet.more.data:'rock+roll'"

    if parts, err := plusSplitter.Split(s); err == nil {
        for _, part := range parts {
            println(part)
        }
    } else {
        println(err.Error())
    }
}

try on go-playground

Disclosure: I am the author of go-andiamo/splitter

How to split on a character only if it's outside of quotes in golang?

2 Answers2