Github user ueshin commented on a diff in the pull request:
https://github.com/apache/spark/pull/18664#discussion_r144248880
--- Diff: python/pyspark/sql/types.py ---
@@ -1619,11 +1619,47 @@ def to_arrow_type(dt):
arrow_type = pa.decimal(dt.precision, dt.scale)
elif type(dt) == StringType:
arrow_type = pa.string()
+ elif type(dt) == DateType:
+ arrow_type = pa.date32()
+ elif type(dt) == TimestampType:
+ arrow_type = pa.timestamp('us', tz='UTC')
else:
raise TypeError("Unsupported type in conversion to Arrow: " +
str(dt))
return arrow_type
+def _check_localize_series_timestamps(s):
+ from pandas.types.common import is_datetime64_dtype
+ # TODO: handle nested timestamps?
+ if is_datetime64_dtype(s.dtype):
+ # TODO: pyarrow.Column.to_pandas keeps data in UTC but removes
timezone
--- End diff --
I'd suspect it's a bug because we expect the both behave the same way and
they shouldn't drop the timezone info.
Anyway, how about creating pandas DataFrame first and split into Series.
I mean as follows:
- python/pyspark/serializers.py
```diff
reader = pa.open_stream(stream)
for batch in reader:
- table = pa.Table.from_batches([batch])
- yield [_check_localize_series_timestamps(c.to_pandas()) for c
in table.itercolumns()]
+ pdf = batch.to_pandas()
+ yield [_check_localize_series_timestamps(c) for _, c in
pdf.iteritems()]
```
- python/pyspark/sql/types.py
```diff
def _check_localize_series_timestamps(s):
- from pandas.types.common import is_datetime64_dtype
+ from pandas.types.common import is_datetime64tz_dtype
# TODO: handle nested timestamps?
- if is_datetime64_dtype(s.dtype):
- # TODO: pyarrow.Column.to_pandas keeps data in UTC but removes
timezone
- return
s.dt.tz_localize('UTC').dt.tz_convert('tzlocal()').dt.tz_localize(None)
+ if is_datetime64tz_dtype(s.dtype):
+ return s.dt.tz_convert('tzlocal()').dt.tz_localize(None)
else:
return s
```
We need to check if this causes performance regression or not.
Btw, when using this patch, we can use `date(1969, 1, 1)` and
`datetime(1969, 1, 1, 1, 1, 1)` for the test you marked as TODO in
`test_vectorized_udf_timestamps`. I'm not exactly sure the reason, though.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
