I'd rather not see ARROW-9223 reverted, if possible. I will put up my
hacked patch to Spark for this so we can test against it if needed, and
could share my branch if anyone else wants to test it locally.
On Sun, Jul 19, 2020 at 7:35 PM Micah Kornfield
wrote:
> I'll spend some time tonight on
I'll spend some time tonight on it and if I can't get round trip working
I'll handle reverting
On Sunday, July 19, 2020, Wes McKinney wrote:
> On Sun, Jul 19, 2020 at 7:33 PM Neal Richardson
> wrote:
> >
> > It sounds like you may have identified a pyarrow bug, which sounds not
> > good,
On Sun, Jul 19, 2020 at 7:33 PM Neal Richardson
wrote:
>
> It sounds like you may have identified a pyarrow bug, which sounds not
> good, though I don't know enough about the broader context to know whether
> this is (1) worse than 0.17 or (2) release blocking. I defer to y'all who
> know better.
It sounds like you may have identified a pyarrow bug, which sounds not
good, though I don't know enough about the broader context to know whether
this is (1) worse than 0.17 or (2) release blocking. I defer to y'all who
know better.
If there are quirks in how Spark handles timezone-naive
Honestly I think reverting is the best option. This change evidently
needs more time to "season" and perhaps this is motivation to enhance
test coverage in a number of places.
On Sun, Jul 19, 2020 at 7:11 PM Wes McKinney wrote:
>
> I am OK with any solution that doesn't delay the production of
I am OK with any solution that doesn't delay the production of the
next RC by more than 24 hours
On Sun, Jul 19, 2020 at 7:08 PM Micah Kornfield wrote:
>
> If I read the example right it looks like constructing from python types
> isn't keeping timezones into in place? I can try make a patch
If I read the example right it looks like constructing from python types
isn't keeping timezones into in place? I can try make a patch that fixes
that tonight or would the preference be to revert my patch (note I think
another bug with a prior bug was fixed in my PR as well)
-Micah
On Sunday,
I put up a PR to revert ARROW-9223. If someone cannot resolve the
problem another way that I recommend applying the reversion and
cutting RC2
https://github.com/apache/arrow/pull/7802
To state the obvious we must verify that this resolves the Spark problem also
On Sun, Jul 19, 2020 at 6:55 PM
I think I see the problem now:
In [40]: parr
Out[40]:
0 {'f0': 1969-12-31 16:00:00-08:00}
1{'f0': 1969-12-31 16:00:00.01-08:00}
2{'f0': 1969-12-31 16:00:00.02-08:00}
dtype: object
In [41]: parr[0]['f0']
Out[41]: datetime.datetime(1969, 12, 31, 16, 0, tzinfo=)
In [42]:
Ah I forgot that this is a "feature" of nanosecond timestamps
In [21]: arr = pa.array([0, 1, 2], type=pa.timestamp('us',
'America/Los_Angeles'))
In [22]: struct_arr = pa.StructArray.from_arrays([arr], names=['f0'])
In [23]: struct_arr.to_pandas()
Out[23]:
0 {'f0': 1969-12-31
There seems to be other broken StructArray stuff
In [14]: arr = pa.array([0, 1, 2], type=pa.timestamp('ns',
'America/Los_Angeles'))
In [15]: struct_arr = pa.StructArray.from_arrays([arr], names=['f0'])
In [16]: struct_arr
Out[16]:
-- is_valid: all not null
-- child 0 type: timestamp[ns,
Well, the problem is that time zones are really finicky comparing
Spark (which uses a localtime interpretation of timestamps without
time zone) and Arrow (which has naive timestamps -- a concept similar
but different from the SQL concept TIMESTAMP WITHOUT TIME ZONE -- and
tz-aware timestamps). So
If it’s a display problem, should it block the release?
Sent from my iPhone
> On Jul 19, 2020, at 3:57 PM, Wes McKinney wrote:
>
> I opened https://issues.apache.org/jira/browse/ARROW-9525 about the
> display problem. My guess is that there are other problems lurking
> here
>
>> On Sun, Jul
I opened https://issues.apache.org/jira/browse/ARROW-9525 about the
display problem. My guess is that there are other problems lurking
here
On Sun, Jul 19, 2020 at 5:54 PM Wes McKinney wrote:
>
> hi Bryan,
>
> This is a display bug
>
> In [6]: arr = pa.array([0, 1, 2], type=pa.timestamp('ns',
>
hi Bryan,
This is a display bug
In [6]: arr = pa.array([0, 1, 2], type=pa.timestamp('ns',
'America/Los_Angeles'))
In [7]: arr.view('int64')
Out[7]:
[
0,
1,
2
]
In [8]: arr
Out[8]:
[
1970-01-01 00:00:00.0,
1970-01-01 00:00:00.1,
1970-01-01 00:00:00.2
]
In
Well this is not good and pretty disappointing given that we had nearly a
month to sort through the implications of Micah’s patch. We should try to
resolve this ASAP
On Sun, Jul 19, 2020 at 5:10 PM Bryan Cutler wrote:
> +0 (non-binding)
>
> I ran verification script for binaries and then
+0 (non-binding)
I ran verification script for binaries and then source, as below, and both
look good
ARROW_TMPDIR=/tmp/arrow-test TEST_DEFAULT=0 TEST_SOURCE=1 TEST_CPP=1
TEST_PYTHON=1 TEST_JAVA=1 TEST_INTEGRATION_CPP=1 TEST_INTEGRATION_JAVA=1
dev/release/verify-release-candidate.sh source 1.0.0
Hi,
It may be a problem of our CMake configuration system.
Could you file this to JIRA?
https://issues.apache.org/jira/projects/ARROW/issues
And could you provide your Portfile?
Thanks,
--
kou
In <43c2cc28-7792-4baa-bdb7-40600812f...@gmail.com>
"Re: Building pyarrow on macOS " on Sun, 19
+1 (binding)
I ran the release verification (source and binary) on Ubuntu 18.04 and
Windows with MSVC.
I experienced a symbol loading issue on macOS [1] but I suspect it's
something environment-specific on the machine. Since we have viable
macOS packages I'm not concerned about this and can
hi folks,
Antoine and I are both on vacation the next two weeks and won't be
doing many code reviews. I'll be back doing some code reviews albeit
at 50% or less capacity the week of July 27. I encourage other
maintainers to step in to help carry the load of getting PRs merged.
If there's anything
This issue has arisen again on a Macports buildbot.
The Apache Arrow cmake system locates the installed libraries and headers, but
then says it can’t find them.
Does anyone know what would cause this build failure? It works on my system,
but not on a build bot:
Build command:
> Executing: cd
Arrow Build Report for Job nightly-2020-07-19-0
All tasks:
https://github.com/ursa-labs/crossbow/branches/all?query=nightly-2020-07-19-0
Failed Tasks:
- conda-linux-gcc-py36-cpu:
URL:
22 matches
Mail list logo