Make zip function available in Jinja

[friendofasquid]

Would like to use zip to loop over some results in a Jinja template

{% for j,k in zip(a, b) %

Currently zip is not available, but would make my code a much cleaner.

[drewbanin]

Thanks for opening this issue @friendofasquid!

I can buy either:

1. Adding zip to the context directly, or
2. Adding it under a namespace

I think if we're only going to add zip, then adding it to the context directly makes sense. If we're instead going to add other Python builtin functions in the future (like set(), dict(), tuple(), or similar) then we'd be better off namespacing the functions to avoid polluting the global namespace.

I think I'd be in favor of adding set at the very least -- maybe we would also benefit from:

• all()
• any()

As far as namespaces go, we already have a builtins namespace, so we could probably add builtins.zip with a single line of code :D

Do you have any thoughts / preferences / ideas? Curious to hear what you think

[friendofasquid]

I don’t know enough about Jinja templating to comment.

My gut would be to have a set of things that aren’t namespaces, mostly because I think dbt is a lot of people’s first time writing code that isn’t SQL.

[github-actions]

This issue has been marked as Stale because it has been open for 180 days with no activity. If you would like the issue to remain open, please remove the stale label or comment on the issue, or it will be closed in 7 days.

[mathcoder3141]

@drewbanin bumping this back up to see if this can be done

[jtcohen6]

I buy the utility of this one, especially zip() + set(). There have been a handful of times I'd wanted to use set logic in particular (example).

I also think the initial reservation may not be as much of an issue, given how we construct builtins: Every method decorated @contextmember gets added to both the global namespace and the builtins namespace. If ever a user wanted to customize the set() macro, they could do it by writing a macro named set() that references builtins.set(). That's (my understanding of) how overrides for ref, source, config work today, and : https://docs.getdbt.com/reference/dbt-jinja-functions/builtins.

The code here can be a little bit tricky, but the basic implementation should look a lot like the tojson method:

• Decorated with @staticmethod + @contextmember
• Perform the Python operation (zip(), set(), etc) and return the value
• Tested in test/unit/test_context.py

I'm going to open this one back up, and mark it as help_wanted (a.k.a. "good second issue," for a community contributor who's comfortable writing a bit of Python)