Setting coerce on a column causes the column to be required when required=False

[adam-porich-sm]

Describe the bug
A clear and concise description of what the bug is.

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of pandera.
• (optional) I have confirmed this bug exists on the main branch of pandera.

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

import polars as pl
import pandera.polars as pa

s = pa.DataFrameSchema(
    {
        "missing_column": pa.Column(int, required=False),
        "int_from_string_column": pa.Column(int, coerce=True),
    }
)

df = pl.DataFrame(
    [
        pl.Series("int_from_string_column", ["1", "2"], dtype=pl.String),
    ]
)

s.validate(df)

s2 = pa.DataFrameSchema(
    {
        "missing_column": pa.Column(int, required=False),
        "int_from_string_column": pa.Column(int),
    },
    coerce=True
)

s2.validate(df)

Expected behavior

I expect both validate calls to succeed

Desktop (please complete the following information):

• OS: Ubuntu 20.04
• Version of Pandera: 0.19.3

Additional context

It appears that the coerce at the DataFrameSchema causes the required=False on the Column to be ignored.

[adam-porich-sm]

Actually here is a more minimal repro

s = pa.DataFrameSchema(
    {
        "missing_column": pa.Column(int, required=False, coerce=True),
    }
)

df = pl.DataFrame()

s.validate(df)

> python3.10 /tmp/repro_test.py
Traceback (most recent call last):
  File "/tmp/repro_test.py", line 12, in <module>
    s.validate(df)
  File "/home/adam.porich/.local/lib/python3.10/site-packages/pandera/api/polars/container.py", line 58, in validate
    output = self.get_backend(check_obj).validate(
  File "/home/adam.porich/.local/lib/python3.10/site-packages/pandera/backends/polars/container.py", line 62, in validate
    check_obj = parser(check_obj, *args)
  File "/home/adam.porich/.local/lib/python3.10/site-packages/pandera/backends/polars/container.py", line 395, in coerce_dtype
    check_obj = self._coerce_dtype_helper(check_obj, schema)
  File "/home/adam.porich/.local/lib/python3.10/site-packages/pandera/backends/polars/container.py", line 454, in _coerce_dtype_helper
    obj = getattr(col_schema.dtype, coerce_fn)(
  File "/home/adam.porich/.local/lib/python3.10/site-packages/pandera/engines/polars_engine.py", line 146, in try_coerce
    lf.collect()
  File "/home/adam.porich/.local/lib/python3.10/site-packages/polars/lazyframe/frame.py", line 1816, in collect
    return wrap_df(ldf.collect(callback))
polars.exceptions.ColumnNotFoundError: missing_column

I was surprised to see coerce=True required that the column exists. I would expect the behaviour of checking if the column exists / is missing first and then if the column exists trying to coerce it