jorisvandenbossche commented on issue #47464: URL: https://github.com/apache/arrow/issues/47464#issuecomment-3257645492
Similarly as in https://github.com/apache/arrow/issues/47460, the issue is the conversion from Arrow->pandas, and not in reading the parquet file. Parquet has a date type, and Arrow as well. And so when reading the Parquet file into an Arrow table using pyarrow, you would see that the type is preserved. You can try: ```python import pyarrow.parquet as pq table = pq.read_table("C:/Users/qnsv2207/Desktop/test_amphi_compare.parquet") table ``` This should show that the "OrderDate" column has a `date32` type. But then pandas does not have a built-in "date" type. Therefore, in the arrow->pandas conversion, pyarrow by default converts its date types into an `object` column with python datetime.date objects. See the documentation about this at https://arrow.apache.org/docs/python/pandas.html#date-types, which also mentions the `date_as_object=False` option you can specify in `to_pandas()` to avoid this conversion to object dtype. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
