jorgecarleitao commented on pull request #7751:
URL: https://github.com/apache/arrow/pull/7751#issuecomment-659041261
> You mean we need Parquet file with the Date64 type?
Yes. That would be the most reliable way to test this. Something equivalent
to the Rust's counter-part of:
```python
import datetime
import pyarrow as pa
import pyarrow.parquet
import os.path
data = [
datetime.datetime(2000, 1, 1, 12, 34, 56, 123456),
datetime.datetime(2000, 1, 1)
]
data32 = pa.array(data, type='date32')
data64 = pa.array(data, type='date64')
table = pyarrow.Table.from_arrays([data32, data64], names=['a', 'b'])
pyarrow.parquet.write_table(table, 'a.parquet')
```
> there is no conversion from Parquet type to Arrow's Date64 type,
therefore the conversion clause may never happen. See here for Rust and here
for C++
Well, I do not know the details, but the fact that when I run
```python
print(table)
print()
print(pyarrow.parquet.read_table('a.parquet'))
```
I get
```
pyarrow.Table
a: date32[day]
b: date64[ms]
pyarrow.Table
a: date32[day]
b: date32[day] <----------- why?
```
does not inspire me much confidence in the current C++/Python implementation
of this particular type. I sense that there is a bug in the current C++'s
implementation of read or write to parquet for this type.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
