jorgecarleitao commented on pull request #7751:
URL: https://github.com/apache/arrow/pull/7751#issuecomment-660445062
I understood that many of the arrow data types are written to parquet with
different types, and are casted to the correct type on read/write.
IMO the reason we cannot
jorgecarleitao commented on pull request #7751:
URL: https://github.com/apache/arrow/pull/7751#issuecomment-660449463
@sunchao , I was able to partially test the date64: I tested the conversion
from i64 to date64 on the converters. Does not cover 100%, since it does not
cover the array_rea
jorgecarleitao commented on pull request #7751:
URL: https://github.com/apache/arrow/pull/7751#issuecomment-660503558
@sunchao, Yeah, re-reading the parquet specification, I agree with you. I
will change this to throw an error for the Data64 branch. Sorry for the
confusion and thank you fo
jorgecarleitao commented on pull request #7751:
URL: https://github.com/apache/arrow/pull/7751#issuecomment-658854948
Thanks @andygrove and @sunchao for taking the time to look at this.
I had to [change an existing
test](https://github.com/apache/arrow/pull/7751/files#diff-fca053527b
jorgecarleitao commented on pull request #7751:
URL: https://github.com/apache/arrow/pull/7751#issuecomment-658906785
> Thanks @jorgecarleitao . Great that the `Date32` is covered. Is it
possible to add a test for `Date64` as well or it is also covered? Ideally we
want to test the error ca
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:
jorgecarleitao commented on pull request #7751:
URL: https://github.com/apache/arrow/pull/7751#issuecomment-659168585
I filed this under a separate issue for Python,C++
[here](https://issues.apache.org/jira/browse/ARROW-9502).
--
jorgecarleitao commented on pull request #7751:
URL: https://github.com/apache/arrow/pull/7751#issuecomment-659971852
Well, then isn't it fair to say that we have no way of testing that we are
correctly reading from date64? we have no way of writing a parquet file with
i64 representing a d