Joining on list columns is not implemented

[shenker]

Description

Related issues: #10747, #12636, #4175, #2869, #13950

df1 = pl.DataFrame({"a": [[1,2,3],[1,2,3],[4,5,6]]})
df2 = pl.DataFrame({"a": [[1,2,3],[4,5,6]], "b": ["x", "y"]})
df1.join(df2, on="a")

gives
PanicException: not implemented.

Note that this error is given even if the join column is a list[cat] or list[enum] column, even though you can compute hashes of those column types and join on the hash. At the very least, polars should automatically do that for you instead of panicking. It is unexpected that group_by works on list[cat] and list[enum] columns but join doesn't.

[tafia]

Similarly:

df1 = pl.DataFrame({"a": [1, 2], "b": [1, 2] })
df2 = df1
df2.join(df1, on=(["a"],))

Also panics (not implemented)

[ritchie46]

If we encounter list columns we should go into the row encoding.