[FEA] support lists::contains on lists of structs.

[revans2]

Is your feature request related to a problem? Please describe.
We would like to extend the lists::contains function to support lists of structs and list of more deeply nested structs, like a struct with structs in it. We are prioritizing structs right now, and this is not intended to cover structs that also contain lists.

Just like with the existing code if the input value is a null or if the search value is a null, then the output should also be a null.

Describe the solution you'd like
Have lists::contains just work.

Describe alternatives you've considered
We could try and hack this together ourselves by picking apart the child columns and then checking for equality for each of them, with and-ing them all together followed by a reduction, but that gets really complicated.

[github-actions]

This issue has been labeled inactive-30d due to no recent activity in the past 30 days. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed. This issue will be labeled inactive-90d if there is no activity in the next 60 days.

[github-actions]

This issue has been labeled inactive-90d due to no recent activity in the past 90 days. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed.

[revans2]

This is still wanted. Not a huge priority though

[ttnghia]

Blocked by hash refactoring. I'm still investigating it.

[jrhemstad]

Blocked by hash refactoring. I'm still investigating it.

This will be solved trivially by #10186

[ttnghia]

@revans2 I think we should better support trivial lists::contain first (i.e., check if each list contains a given scalar), as we did for the general cudf::contain before. General support for checking a list containing a set a scalars is more difficult, is currently blocked by hash refactoring. We will implement it later if there is customer request.