The results of selecting columns with regular expressions in the context of groupby are wrong on Polars 0.18.7.

[Teamon9161]

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

df = pl.DataFrame({'g':[3], 'abc': [1],'abcd': [3]})

print(df.groupby('g').agg(pl.col('^ab.*$')))

Issue description

The result of the above code is correct in Polars 0.18.6 and below, but the result is incorrect in version 0.18.7.
Correct output:

shape: (1, 3)
┌─────┬───────────┬───────────┐
│ g   ┆ abc       ┆ abcd      │
│ --- ┆ ---       ┆ ---       │
│ i64 ┆ list[i64] ┆ list[i64] │
╞═════╪═══════════╪═══════════╡
│ 3   ┆ [1]       ┆ [3]       │
└─────┴───────────┴───────────┘

Wrong output

shape: (1, 1)
┌─────┐
│ g   │
│ --- │
│ i64 │
╞═════╡
│ 3   │
└─────┘

When I run df.select(pl.col('^ab.*$')) on Polars 0.18.7, the result is also correct.

shape: (1, 2)
┌─────┬──────┐
│ abc ┆ abcd │
│ --- ┆ ---  │
│ i64 ┆ i64  │
╞═════╪══════╡
│ 1   ┆ 3    │
└─────┴──────┘

Expected behavior

See above

Installed versions

polars 0.18.6

--------Version info---------
Polars:              0.18.6
Index type:          UInt32
Platform:            Windows-10-10.0.22621-SP0
Python:              3.10.11 | packaged by Anaconda, Inc. | (main, Apr 20 2023, 18:56:50) [MSC v.1916 64 bit (AMD64)]

----Optional dependencies----
adbc_driver_sqlite:  <not installed>
connectorx:          0.3.1
deltalake:           0.9.0
fsspec:              2023.5.0
matplotlib:          3.7.1
numpy:               1.24.3
pandas:              1.5.3
pyarrow:             12.0.0
pydantic:            <not installed>
sqlalchemy:          2.0.15
xlsx2csv:            0.8.1
xlsxwriter:          3.1.2

polars 0.18.7

--------Version info---------
Polars:              0.18.7
Index type:          UInt32
Platform:            Windows-10-10.0.22621-SP0
Python:              3.10.11 | packaged by Anaconda, Inc. | (main, Apr 20 2023, 18:56:50) [MSC v.1916 64 bit (AMD64)]

----Optional dependencies----
adbc_driver_sqlite:  <not installed>
connectorx:          0.3.1
deltalake:           0.9.0
fsspec:              2023.5.0
matplotlib:          3.7.1
numpy:               1.24.3
pandas:              1.5.3
pyarrow:             12.0.0
pydantic:            <not installed>
sqlalchemy:          2.0.15
xlsx2csv:            0.8.1
xlsxwriter:          3.1.2

[cmdlineluser]

Looks like it was introduced when fixing .exclude() #9827

It appears that the block on L148 is never entered.

[/Users/user/git/polars/polars/polars-lazy/polars-plan/src/logical_plan/projection.rs:150] exclude = {
    "g",
}
[/Users/user/git/polars/polars/polars-lazy/polars-plan/src/logical_plan/projection.rs:153] e = col("^ab.*$")

polars/polars/polars-lazy/polars-plan/src/logical_plan/projection.rs

Lines 145 to 148 in baad248

	// iterate until we find the Exclude node
	// we remove that node from the expression
	for e in expr.into_iter() {
	if let Expr::Exclude(e, _) = e {

So it looks like groupby keys are added to exclude meaning it's possible for expr to not contain any exclude nodes?