REGEX operator only matches characters at the beginning of a string

[rolodato]

REGEX is implemented using Pattern.match, which only checks the beginning of the search string:

If zero or more characters at the beginning of string match this regular expression, return a corresponding Match

This means a regular expression like world will not match a trait value of hello-world.

We should be using Pattern.search instead: https://docs.python.org/3.11/library/re.html#re.Pattern.search

[khvn26]

I'm leaning towards closing this issue as a wontfix:

1. Such a change would be breaking for all SaaS customers.
2. re.search is significantly less performant.
3. A pattern like .*world should cater for the described use case, unless I'm missing something.

[rolodato]

I can confirm using a pattern like .*world.* works, thanks for that suggestion!

There is probably a better way than using Pattern.search, it was the first thing that came to mind :)

I would argue this is a bugfix and not a breaking change. We don't document this specifically, but I would say the expectation for "matches regex" is essentially the same way grep behaves. grep world does match hello-world-hello, and does not require adding anything else around world to match.

We have this assumption built-in to every SDK that I've checked so far.

Node.js

> 'hello-world-hello'.match(/world/)
[ 'world', index: 6, input: 'hello-world-hello', groups: undefined ]

Ruby

irb(main):001> 'hello-world-hello'.match?('world')
=> true

Golang

package main

import "regexp"

func main() {
	matched, err := regexp.Match(`world`, []byte(`hello-world-hello`))
	println(matched, err)
	// true, nil
}