Multi index dataframes with datetime level cannot drop invalid rows

[rorymcstay]

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 master 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

We use raw eval on interpretting string tuples. This fails under datetime types serialised as Timestamp(...) in a string.

Traceback (most recent call last):
  File "/opt/atlassian/pipelines/agent/build/instrument-svc/instrument_svc/test/downloader_persistence/test_arctic_cache.py", line 265, in test_validate_views
    cli.validate(start_date=start_date, end_date=end_date,
  File "/opt/atlassian/pipelines/agent/build/instrument-svc/instrument_svc/downloader/cli.py", line 188, in validate
    good_data, report = view.validate(
  File "/opt/atlassian/pipelines/agent/build/instrument-svc/instrument_svc/downloader/view.py", line 303, in validate
    validated_df = schema.validate(data, lazy=True, inplace=True)
  File "/opt/atlassian/pipelines/agent/build/venv/lib/python3.8/site-packages/pandera/api/pandas/model.py", line 306, in validate
    cls.to_schema().validate(
  File "/opt/atlassian/pipelines/agent/build/venv/lib/python3.8/site-packages/pandera/api/pandas/container.py", line 375, in validate
    return self._validate(
  File "/opt/atlassian/pipelines/agent/build/venv/lib/python3.8/site-packages/pandera/api/pandas/container.py", line 404, in _validate
    return self.get_backend(check_obj).validate(
  File "/opt/atlassian/pipelines/agent/build/venv/lib/python3.8/site-packages/pandera/backends/pandas/container.py", line 112, in validate
    check_obj = self.drop_invalid_rows(check_obj, error_handler)
  File "/opt/atlassian/pipelines/agent/build/venv/lib/python3.8/site-packages/pandera/backends/pandas/base.py", line 170, in drop_invalid_rows
    index_tuples = err.failure_cases["index"].apply(eval)
  File "/opt/atlassian/pipelines/agent/build/venv/lib/python3.8/site-packages/pandas/core/series.py", line 4771, in apply
    return SeriesApply(self, func, convert_dtype, args, kwargs).apply()
  File "/opt/atlassian/pipelines/agent/build/venv/lib/python3.8/site-packages/pandas/core/apply.py", line 1123, in apply
    return self.apply_standard()
  File "/opt/atlassian/pipelines/agent/build/venv/lib/python3.8/site-packages/pandas/core/apply.py", line 1174, in apply_standard
    mapped = lib.map_infer(
  File "pandas/_libs/lib.pyx", line 2924, in pandas._libs.lib.map_infer
  File "<string>", line 1, in <module>
NameError: name 'Timestamp' is not defined

Easy work around would be to pass {'Timestamp': pd.Timestamp} in globals to eval

Expected behavior

Invalid rows are dropped

Additional context

Add any other context about the problem here.

[rorymcstay]

I have a fix for this, Will raise a PR today

[rorymcstay]

It would also be useful to have the SchemaErrors object in the case that there are invalid rows, and we opt to drop them.

I have a workflow whereby I run a validation, dropping invalid rows, and persist the error report of the invalid observations. This way I can track data quality over time.

It would be useful to have that. I can add after or as part of the fix to this.

[cosmicBboy]

Sounds good! Were you thinking the errors object would be in the DataFrame.pandera accessor?