MINOR: [Python] fix pandas_compat.py

[Piong1997]

There is a bug that when column dtype is np.bytes，it will goto the final branch and run level=level.astype(dtype)

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

[github-actions]

Thanks for opening a pull request!

If this is not a minor PR. Could you open an issue for this pull request on GitHub? https://github.com/apache/arrow/issues/new/choose

Opening GitHub issues ahead of time contributes to the Openness of the Apache Arrow project.

Then could you also rename the pull request title in the following format?

GH-${GITHUB_ISSUE_ID}: [${COMPONENT}] ${SUMMARY}

or

MINOR: [${COMPONENT}] ${SUMMARY}

In the case of PARQUET issues on JIRA the title also supports:

PARQUET-${JIRA_ISSUE_ID}: [${COMPONENT}] ${SUMMARY}

See also:

• Other pull requests
• Contribution Guidelines - Contributing Overview

[jorisvandenbossche]

@Piong1997 do you have a reproducible example that shows the bug? (it would be good to add a test for it)

[Piong1997]

@Piong1997 do you have a reproducible example that shows the bug? (it would be good to add a test for it)

Sure, the following code reproduce this bug:

import pandas as pd # pd.__version__ is 2.0.3
import pyarrow # pyarrow.__version__ is 17.0.0

df = pd.DataFrame([[0.1, 0.2], [0.3, 0.4]], columns=[b'col1', b'col2'])
table = pyarrow.Table.from_pandas(df)
table.to_pandas() # error occurred

Note that this bug can only be reproduced in pandas 2.0 or above.
This bug was introduced in a certain version of pyarrow, pyarrow version 6.0 can work normally.