Inconsistent behavior of n_nunique for series and dataframe

[vlulla]

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

df = pl.DataFrame({"list-of-structs":[[{"a":1,"b":2},{"a":10,"b":20}],
                                                               [{"a":1,"b":2},{"a":10,"b":20}],
                                                               [{"a":3,"b":4}]]
                                })
print(df.select('list-of-structs').n_unique()," <== works") # works
print(type(df.select('list-of-structs'))) # polars.dataframe.frame.DataFrame

print(df['list-of-structs'].n_unique()) # InvalidOperationError: `n_unique` operation not supported for dtype `list[struct[2]]`
print(type(df['list-of-structs'])) # polars.series.series.Series

I am unsure why the series version of n_unique will not work whereas the dataframe version will work. I tried looking up this information in the documentation but was unable to find anything substantive.

Log output

No response

Issue description

Inconsistent behavior of n_unique for a series of list-of-structs.

Expected behavior

I believe that df.select('list-of-structs').n_unique() and df['list-of-structs'].n_unique() both ought to work and give the same answer. Currently the series version raises an InvalidOperationError.

Installed versions

--------Version info---------
Polars:              1.17.1
Index type:          UInt32
Platform:            macOS-14.4.1-arm64-arm-64bit-Mach-O
Python:              3.13.1 | packaged by conda-forge | (main, Dec  5 2024, 21:09:18) [Clang 18.1.8 ]
LTS CPU:             False

----Optional dependencies----
adbc_driver_manager  <not installed>
altair               <not installed>
boto3                <not installed>
cloudpickle          <not installed>
connectorx           <not installed>
deltalake            <not installed>
fastexcel            <not installed>
fsspec               <not installed>
gevent               <not installed>
google.auth          2.37.0
great_tables         <not installed>
matplotlib           3.10.0
nest_asyncio         1.6.0
numpy                2.2.0
openpyxl             3.1.5
pandas               2.2.3
pyarrow              18.1.0
pydantic             <not installed>
pyiceberg            <not installed>
sqlalchemy           2.0.36
torch                <not installed>
xlsx2csv             <not installed>
xlsxwriter           <not installed>

[cmdlineluser]

It seems even with expressions, it errors.

df.select(pl.col("list-of-structs").n_unique())
# InvalidOperationError: `n_unique` operation not supported for dtype `list[struct[2]]`

Not sure if it is due to this check:

polars/crates/polars-core/src/series/implementations/list.rs

Line 203 in 676f10d

if !self.inner_dtype().is_numeric() {

It seems the DataFrame method wraps it in a pl.struct call.

polars/py-polars/polars/dataframe/frame.py

Line 10210 in 676f10d

expr = F.struct(struct_fields)

df.select(pl.struct("list-of-structs").n_unique())
# shape: (1, 1)
# ┌─────────────────┐
# │ list-of-structs │
# │ ---             │
# │ u32             │
# ╞═════════════════╡
# │ 2               │
# └─────────────────┘