str.to_time() does not work if all values are null

[lonestike]

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(
    {"timestamp": [None, None], "other": ["text1", "text2"]}, 
    schema_overrides={"timestamp": pl.String},
)

# does not work:
df.with_columns(pl.col("timestamp").str.to_time())

# does work
df.with_columns(pl.col("timestamp").str.to_time("%H"))

Issue description

Hello,
When all the values are null in a string column, polars can't parse them to time unless we specify a format (whatever the format is, it works)

maybe linked to #22370 ?

Expected behavior

expect df.with_columns(pl.col("timestamp").str.to_time()) to work when all the values are null in a str column

Installed versions

\--------Version info---------
Polars:              1.29.0
Index type:          UInt32
Platform:            Linux-3.10.0-1160.108.1.el7.x86\_64-x86\_64-with-glibc2.17
Python:              3.10.4 (main, Mar 31 2022, 08:41:55) \[GCC 7.5.0]
LTS CPU:             False

\----Optional dependencies----
Azure CLI            <not installed>
adbc\_driver\_manager  1.6.0
altair               <not installed>
azure.identity       1.21.0
boto3                1.35.42
cloudpickle          3.0.0
connectorx           <not installed>
deltalake            0.25.5
fastexcel            <not installed>
fsspec               2025.3.2
gevent               <not installed>
google.auth          2.38.0
great\_tables         <not installed>
matplotlib           3.9.2
numpy                2.0.0
openpyxl             3.1.5
pandas               2.2.3
polars\_cloud         <not installed>
pyarrow              17.0.0
pydantic             2.11.4
pyiceberg            0.9.0
sqlalchemy           2.0.35
torch                <not installed>
xlsx2csv             <not installed>
xlsxwriter           3.2.0