join_where ColumnNotFoundError if predicate only uses columns from one side

[cmdlineluser]

Checks

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of Polars.

Reproducible example

import polars as pl

left  = pl.DataFrame({"a": [1, 2]}).with_row_index()
right = pl.DataFrame({"b": [1, 2]}).with_row_index()

left.join_where(right, pl.col.index >= pl.col.a)
# ColumnNotFoundError: could not find column a_right in the right table during join operation

Log output

No response

Issue description

It's a nonsense example, but the a_right seems to suggest the order of col() matters in this case.

Reversing the order does produce a result:

>>> left.join_where(right, pl.col.a <= pl.col.index)
# shape: (1, 4)
# ┌───────┬─────┬─────────────┬─────┐
# │ index ┆ a   ┆ index_right ┆ b   │
# │ ---   ┆ --- ┆ ---         ┆ --- │
# │ u32   ┆ i64 ┆ u32         ┆ i64 │
# ╞═══════╪═════╪═════════════╪═════╡
# │ 0     ┆ 1   ┆ 1           ┆ 2   │
# └───────┴─────┴─────────────┴─────┘

Which looks like it actually compared a and index_right?

It seems like a predicate that only refers to columns from one side should be considered Invalid?

Expected behavior

I'm not sure.

Installed versions

--------Version info---------
Polars:               1.7.1
Index type:           UInt32
Platform:             macOS-13.6.1-arm64-arm-64bit
Python:               3.12.2 (main, Feb  6 2024, 20:19:44) [Clang 15.0.0 (clang-1500.1.0.2.5)]

----Optional dependencies----
adbc_driver_manager:  <not installed>
cloudpickle:          <not installed>
connectorx:           <not installed>
deltalake:            <not installed>
fastexcel:            <not installed>
fsspec:               <not installed>
gevent:               <not installed>
great_tables:         <not installed>
hvplot:               <not installed>
matplotlib:           <not installed>
nest_asyncio:         <not installed>
numpy:                1.26.4
openpyxl:             <not installed>
pandas:               2.2.1
pyarrow:              15.0.2
pydantic:             <not installed>
pyiceberg:            <not installed>
sqlalchemy:           <not installed>
torch:                <not installed>
xlsx2csv:             <not installed>
xlsxwriter:           <not installed>

[ritchie46]

It seems like a predicate that only refers to columns from one side should be considered Invalid?

Yes, it should. We should be able to catch that.