Regex that match multiple format to extract aws accountid

I want to use one regex to extract the aws account id for two types of format

arn:aws:ec2:us-west-2:1234567890:instance/i-0a89fdbedc3f1d76a
arn:aws:iam::1234567890:instance-profile/aws-elasticbeanstalk-ec2-role

I.e., the regex should be able to match 1234567890 in both lines.

I know how to create regex for each format, but want to have one single regex if possible.

I'd use the following regexp: arn:aws:(?:ec2|iam):[^:]*:([0-9]+):
The prefix is arn:aws:, followed by either ec2 or iam and one may-be-empty segment. After the prefix there must be a non-empty group of digits. It is exactly the group with the instance number.

Example https://go.dev/play/p/mP6VyIYa3v_N

package main
import (
	"fmt"
	"regexp"
)
const (
	s1 = "arn:aws:ec2:us-west-2:1234567890:instance/i-0a89fdbedc3f1d76a"
	s2 = "arn:aws:iam::1234567890:instance-profile/aws-elasticbeanstalk-ec2-role"
)
func main() {
	reStr := `arn:aws:(?:ec2|iam):[^:]*:([0-9]+):`
	re, err := regexp.Compile(reStr)
	if err != nil {
		panic(err)
	}
	m1 := re.FindStringSubmatch(s1)
	fmt.Println("User id in ec2: ", m1[1])
	m2 := re.FindStringSubmatch(s2)
	fmt.Println("User id in iam: ", m2[1])
}

This regexp uses non-capturing group (?:ec2|iam). The submatch that corresponds to this group is not included in the array returned by regexp.FindStringSubmatch. This is done for conveniece only: the array returned by FindStringSubmatch has only 2 elements - the full match and the match for the group ([0-9]+).

You can use full group (ec2|iam). In this case the User ID group will have index 2 in the array of matches returned by FindStringSubmatch.

Dont need RegExp, just pick out the fifth column:

package aws
import (
   "strconv"
   "strings"
)
func extract(s string) (int, error) {
   split := strings.SplitN(s, ":", 6)
   return strconv.Atoi(split[4])
}

https://godocs.io/strings#SplitN