feat: support lambda left, right: (left.col == right.col) | (left.col.isnull() & right.col.isnull()) as predicate in joins

[NickCrews]

Is your feature request related to a problem?

I have several tables. I want to join them where several columns are equal or both null.

I currently do

import ibis
a = ibis.memtable({"x": [1,2,None], "y": ["a", "b", None]})
b = ibis.memtable({"x": [None, 2,None], "y": ["a", "b", None]})
c = ibis.memtable({"x": [3, 2,None], "y": ["a", "b", None]})

def equalish(left, right, col):
    return (left[col] == right[col]) | (left[col].isnull() & right[col].isnull())

ab = a.join(b, [equalish(a, b, "x"), equalish(a, b, "y")])
bc = a.join(b, [equalish(b, c, "x"), equalish(b, c, "y")])

This is annoying because I need to have two slightly different lists of join predicates. I want to be able to make this DRY, and only define the join predicates in one place. eg I want to be able to do

def equalish(col: str):
    def pred(left: ibis.Table, right: ibis.Table):
        return (left[col] == right[col]) | (left[col].isnull() & right[col].isnull())
   return pred

preds = [equalish("x"), equalish("y")]

ab = a.join(b, preds)
bc = a.join(b, preds)

What is the motivation behind your request?

No response

Describe the solution you'd like

join() should also support:

• Callable[[ir.Table, ir.Table], ] (eg equalish("x"))
• Iterable[Callable[[ir.Table, ir.Table], ]] (eg [equalish("x"), equalish("y")]

What version of ibis are you running?

main

What backend(s) are you using, if any?

No response

Code of Conduct

• I agree to follow this project's Code of Conduct

[NickCrews]

This feature is included as part of #10703, I think we should discuss it there where these is other context