String panic on compare int to str

[paddymul]

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

cleaned  = pl.DataFrame({'a': [  10 , 20]})
original = pl.DataFrame({'a': [ "10", 20]})

expected = pl.DataFrame(
    {'a'     : [  10 ,   20],
     'a_orig': [ "10", None]})

cleaned.select(
    pl.col('a'),
    pl.when((original['a'] == cleaned['a']).eq(0)).then(None).otherwise(cleaned['a']).alias("a_orig"))

Log output

-traceback
thread '<unnamed>' panicked at crates/polars-core/src/series/arithmetic/borrowed.rs:426:44:
data types don't match: InvalidOperation(ErrString("sub operation not supported for dtypes `str` and `str`"))
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

---------------------------------------------------------------------------
PanicException                            Traceback (most recent call last)
Cell In[119], line 10
      2 original = pl.DataFrame({'a': [ "10", 20]})
      4 expected = pl.DataFrame(
      5     {'a'     : [  10 ,   20],
      6      'a_orig': [ "10", None]})
      8 cleaned.select(
      9     pl.col('a'),
---> 10     pl.when((original['a'] - cleaned['a']).eq(0)).then(None).otherwise(cleaned['a']).alias("a_orig"))

File ~/anaconda3/envs/buckaroo-dev-5/lib/python3.11/site-packages/polars/series/series.py:1068, in Series.__sub__(self, other)
   1066 if isinstance(other, pl.Expr):
   1067     return F.lit(self) - other
-> 1068 return self._arithmetic(other, "sub", "sub_<>")

File ~/anaconda3/envs/buckaroo-dev-5/lib/python3.11/site-packages/polars/series/series.py:1005, in Series._arithmetic(self, other, op_s, op_ffi)
   1002     other = pl.Series("", [None])
   1004 if isinstance(other, Series):
-> 1005     return self._from_pyseries(getattr(self._s, op_s)(other._s))
   1006 elif _check_for_numpy(other) and isinstance(other, np.ndarray):
   1007     return self._from_pyseries(getattr(self._s, op_s)(Series(other)._s))

PanicException: data types don't match: InvalidOperation(ErrString("sub operation not supported for dtypes `str` and `str`"))

Issue description

I'm not sure if this is an issue. I think panics are different than other errors. Also the error message is confusing.
```"sub operation not supported for dtypes str and `str`"

I am subtracting  an int column from a string column, not `str` from `str`

### Expected behavior

I guess I would expect an error, but I'm not sure if this is the right type of error.  Also the message could be clearer.

I want to be able to get the dataframe in `expected`.

I tried with objects, and it too failed.

```python
cleaned  = pl.DataFrame({'a': pl.Series([  10 , 20], dtype=pl.Object)})
original = pl.DataFrame({'a': pl.Series([ "10", 20], dtype=pl.Object)})

expected = pl.DataFrame(
    {'a'     : [  10 ,   20],
     'a_orig': [ "10", None]})

cleaned.select(
    pl.col('a'),
    pl.when((original['a'] == cleaned['a']).eq(True)).then(None).otherwise(cleaned['a']).alias("a_orig"))

Installed versions

Polars:               0.20.7
Index type:           UInt32
Platform:             macOS-13.6.1-arm64-arm-64bit
Python:               3.11.7 (main, Dec 15 2023, 12:09:56) [Clang 14.0.6 ]

----Optional dependencies----
adbc_driver_manager:  <not installed>
cloudpickle:          <not installed>
connectorx:           <not installed>
deltalake:            <not installed>
fsspec:               <not installed>
gevent:               <not installed>
hvplot:               <not installed>
matplotlib:           3.8.2
numpy:                1.26.3
openpyxl:             <not installed>
pandas:               2.1.4
pyarrow:              11.0.0
pydantic:             2.5.3
pyiceberg:            <not installed>
pyxlsb:               <not installed>
sqlalchemy:           <not installed>
xlsx2csv:             <not installed>
xlsxwriter:           <not installed>

[mcrumiller]

On the current main branch this gives a ComputeError and does not panic:

polars.exceptions.ComputeError: cannot compare string with numeric data

[ritchie46]

Then I think it is the proper behavior on main. We can add a test and close.

[mcrumiller]

We already have a test for that here. We can close.