map_elements silently swallows Exceptions when return_dtype= is set

[cmdlineluser]

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

pl.select(
    pl.lit(1).map_elements(lambda x: 
        I | DO | NOT | EXIST,
        return_dtype=pl.String
    )
)

# shape: (1, 1)
# ┌─────────┐
# │ literal │
# │ ---     │
# │ str     │
# ╞═════════╡
# │ null    │
# └─────────┘

Log output

No response

Issue description

I previously encountered this while debugging a related issue: #14821 (comment) (some notes there)

Breaking out as a separate issue as I just ran into a confusing few minutes due to it.

Expected behavior

.map_batches propagates the error.

# ComputeError: NameError: name 'I' is not defined

Installed versions

--------Version info---------
Polars:               1.9.0
Index type:           UInt32
Platform:             macOS-13.6.1-arm64-arm-64bit
Python:               3.12.2 (main, Feb  6 2024, 20:19:44) [Clang 15.0.0 (clang-1500.1.0.2.5)]

----Optional dependencies----
adbc_driver_manager:  <not installed>
cloudpickle:          <not installed>
connectorx:           <not installed>
deltalake:            <not installed>
fastexcel:            <not installed>
fsspec:               <not installed>
gevent:               <not installed>
great_tables:         <not installed>
hvplot:               <not installed>
matplotlib:           <not installed>
nest_asyncio:         <not installed>
numpy:                1.26.4
openpyxl:             <not installed>
pandas:               2.2.1
pyarrow:              15.0.2
pydantic:             <not installed>
pyiceberg:            <not installed>
sqlalchemy:           <not installed>
torch:                <not installed>
xlsx2csv:             <not installed>
xlsxwriter:           <not installed>

[DavideCanton]

I was going to open a separate issue, but maybe it's better to add what I noticed to this, since the source of the problem seems the same.

This behavior is quite unfortunate when it comes to handling warning filters.

This code:

import warnings
import polars as pl


def warns(x):
    warnings.warn("deprecated", DeprecationWarning)
    return x


df = pl.DataFrame({"a": [1, 2, 3, 4, 5]})
print(df.select(pl.col("a").map_elements(warns, return_dtype=pl.Int64)))

If ran with python foo.py prints:

/home/canton/source/prove/foo.py:6: DeprecationWarning: deprecated
  warnings.warn("deprecated", DeprecationWarning)
shape: (5, 1)
┌─────┐
│ a   │
│ --- │
│ i64 │
╞═════╡
│ 1   │
│ 2   │
│ 3   │
│ 4   │
│ 5   │
└─────┘

while if ran with python -W error foo.py prints:

shape: (5, 1)
┌──────┐
│ a    │
│ ---  │
│ i64  │
╞══════╡
│ null │
│ null │
│ null │
│ null │
│ null │
└──────┘

Removing the return_dtype parameter restores the correct behavior.

[maitreyakv]

I have noticed that adding in the skip_nulls=True in map_elements option seems to resolve the behavior, and the exception is properly raised.