SQL: performing a join that contains both an explicit, and implicit join fails if trying to select table in aggregation

[universalmind303]

Describe the bug

So this is very oddly specific, but it is pretty common in a lot of the tpc-ds queries.

select count(tbl_3.val3)
from tbl_1 
join tbl_2 on tbl_1.id = tbl_2.id, tbl_3
where tbl_1.id = tbl_3.id
group by tbl_2.val2

To Reproduce

tbl_1 = daft.from_pydict({
    'id': [1,2,3],
    'val': ['a','b','c']
})

tbl_2 = daft.from_pydict({
    'id': [1,2,4], 
    'val2': ['x','y','z']
})

tbl_3 = daft.from_pydict({
    'id': [1,2,5],
    'val3': ['m','n','o'] 
})

tbl_4 = daft.from_pydict({
    'id': [1,3,6],
    'val4': ['p','q','r']
})
daft.sql("""
select count(tbl_3.val3), tbl_2.val2
from tbl_1 
join tbl_2 on tbl_1.id = tbl_2.id, tbl_3
where tbl_1.id = tbl_3.id
group by tbl_2.val2

""").collect()

results in InvalidSQLException: Table not found: tbl_2

Expected behavior

should have the same output as other sql engines
duckdb:

┌───────────────────┬─────────┐
│ count(tbl_3.val3) │  val2   │
│       int64       │ varchar │
├───────────────────┼─────────┤
│                 1 │ y       │
│                 1 │ x       │
└───────────────────┴─────────┘

Component(s)

SQL

Additional context

No response