Coalescing outer join panics (and/or loses) columns from right frame if join keys expressions have overlapping names

[wence-]

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, 3], "b": [3, 4, 5], "c": [5, 6, 7]})
right = pl.DataFrame({"a": [2, 3, 4], "c": [4, 5, 6]})
left.join(right, on=[pl.col("a")], how="outer_coalesce")
# shape: (4, 4)
# ┌─────┬──────┬──────┬─────────┐
# │ a   ┆ b    ┆ c    ┆ c_right │
# │ --- ┆ ---  ┆ ---  ┆ ---     │
# │ i64 ┆ i64  ┆ i64  ┆ i64     │
# ╞═════╪══════╪══════╪═════════╡
# │ 2   ┆ 4    ┆ 6    ┆ 4       │
# │ 3   ┆ 5    ┆ 7    ┆ 5       │
# │ 4   ┆ null ┆ null ┆ 6       │
# │ 1   ┆ 3    ┆ 5    ┆ null    │
# └─────┴──────┴──────┴─────────┘

# nonsensical, but ok
left.join(right, on=[pl.col("a"), pl.col("a")], how="outer_coalesce")
# shape: (4, 3)
# ┌─────┬──────┬──────┐
# │ a   ┆ b    ┆ c    │
# │ --- ┆ ---  ┆ ---  │
# │ i64 ┆ i64  ┆ i64  │
# ╞═════╪══════╪══════╡
# │ 2   ┆ 4    ┆ 6    │
# │ 3   ┆ 5    ┆ 7    │
# │ 4   ┆ null ┆ null │
# │ 1   ┆ 3    ┆ 5    │
# └─────┴──────┴──────┘

# even more
left.join(right, on=[pl.col("a"), pl.col("a"), pl.col("a")], how="outer_coalesce")
# thread '<unnamed>' panicked at crates/polars-ops/src/frame/join/general.rs:90:25:
# removal index (is 3) should be < len (is 3)

Log output

run JoinExec
join parallel: true
OUTER join dataframes finished
run JoinExec
join parallel: true
OUTER join dataframes finished
run JoinExec
join parallel: true

Issue description

Looks like coalescing outer join just attempts to eat as many columns from the right dataframe as there are key columns in the join.

Expected behavior

I would expect all three of these (the latter two being odd) mathematically equivalent join expressions to give me the same result.

Or, complain that we're going to produce overlapping output key names.

Installed versions

--------Version info---------
Polars:               0.20.26
Index type:           UInt32
Platform:             Linux-6.5.0-35-generic-x86_64-with-glibc2.35
Python:               3.10.14 | packaged by conda-forge | (main, Mar 20 2024, 12:45:18) [GCC 12.3.0]

----Optional dependencies----
adbc_driver_manager:  0.11.0
cloudpickle:          3.0.0
connectorx:           0.3.3
deltalake:            0.17.4
fastexcel:            0.10.4
fsspec:               2024.3.1
gevent:               24.2.1
hvplot:               0.10.0
matplotlib:           3.8.4
nest_asyncio:         1.6.0
numpy:                1.26.4
openpyxl:             3.1.2
pandas:               2.2.2
pyarrow:              16.0.0
pydantic:             2.7.1
pyiceberg:            <not installed>
pyxlsb:               1.0.10
sqlalchemy:           2.0.30
torch:                2.3.0.post300
xlsx2csv:             0.8.2
xlsxwriter:           3.2.0

[ritchie46]

Does this happen if we don't join on twice the same name? We should raise as it doesn't make sense to join on duplicate columns.

[wence-]

Thanks!