Cannot use regular expressions in `pl.col` within an `over`

[avimallu]

Checks

• [X] I have checked that this issue has not already been reported.

• [X] I have confirmed this bug exists on the latest version of Polars.

Reproducible example

import polars as pl

(
    pl.DataFrame({
        "grp_1": [1, 1, 2, 2, 2, 3, 3, 3, 3],
        "grp_2": [2, 1, 2, 1, 2, 2, 1, 1, 2],
        "value": [1, 2, 3, 1, 3, 1, 4, 2, 1]})
    .with_columns(
        pl.col("value").sum().over(pl.col("^grp.*$"))
    )
)

Log output

ComputeError: The name: 'value' passed to `LazyFrame.with_columns` is duplicate

Error originated just after this operation:
DF ["grp_1", "grp_2", "value"]; PROJECT */3 COLUMNS; SELECTION: "None"

Issue description

I cannot use regular expressions within an over call to perform operations - it looks like this gets expanded into multiple columns, resulting in duplicate output column names.

Expected behavior

Identical behaviour to:

(
    pl.DataFrame({
        "grp_1": [1, 1, 2, 2, 2, 3, 3, 3, 3],
        "grp_2": [2, 1, 2, 1, 2, 2, 1, 1, 2],
        "value": [1, 2, 3, 1, 3, 1, 4, 2, 1]})
    .with_columns(
        pl.col("value").sum().over("grp_1", "grp_2")
    )
)

Installed versions

``` --------Version info--------- Polars: 0.19.19 Index type: UInt32 Platform: macOS-14.1.2-arm64-arm-64bit Python: 3.10.8 | packaged by conda-forge | (main, Nov 22 2022, 08:25:13) [Clang 14.0.6 ] ----Optional dependencies---- adbc_driver_manager: 0.5.1 cloudpickle: 2.2.1 connectorx: 0.3.1 deltalake: 0.10.0 fsspec: 2023.9.2 gevent: matplotlib: 3.7.2 numpy: 1.24.3 openpyxl: 3.0.10 pandas: 2.1.1 pyarrow: 11.0.0 pydantic: 2.0.1 pyiceberg: pyxlsb: sqlalchemy: 2.0.21 xlsx2csv: 0.8.1 xlsxwriter: 3.0.9 ```

[cmdlineluser]

Using a struct works: pl.struct("^grp.*$")

[avimallu]

Thanks! I wasn't aware that regular expressions could be used in structs as well.

[cmdlineluser]

Yeah, it almost seems like whatever is passed to over should be implicitly wrapped in a struct?

e.g. I would have thought this would work:

df.with_columns(
   pl.col("value").sum().over("^grp.*$")
)
# ComputeError: The name: 'value' passed to `LazyFrame.with_columns` is duplicate

[datenzauberai]

Using pl.exclude also seems to be affected:

# over with multiple explicit columns: ok
df.with_columns(
    pl.col("value").min().over(["grp_1", "grp_2"]).alias("minimum")
)

# over with single column via exclude: ok
df.drop("grp_2").with_columns(
    pl.col("value").min().over(pl.exclude("value")).alias("minimum")
)

# over with multiple columns via exclude: "the name: 'minimum' passed to `LazyFrame.with_columns` is duplicate"
df.with_columns(
    pl.col("value").min().over((pl.exclude("value"))).alias("minimum")
)

[ritchie46]

This is expected: #19681