[jira] [Created] (ARROW-1339) [C++] Use boost::filesystem for handling of platform-specific file path encodings
Wes McKinney created ARROW-1339: --- Summary: [C++] Use boost::filesystem for handling of platform-specific file path encodings Key: ARROW-1339 URL: https://issues.apache.org/jira/browse/ARROW-1339 Project: Apache Arrow Issue Type: Improvement Components: C++ Reporter: Wes McKinney Fix For: 0.7.0 -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (ARROW-1338) [Python] Investigate non-deterministic core dump on Python 2.7, Travis CI builds
Wes McKinney created ARROW-1338: --- Summary: [Python] Investigate non-deterministic core dump on Python 2.7, Travis CI builds Key: ARROW-1338 URL: https://issues.apache.org/jira/browse/ARROW-1338 Project: Apache Arrow Issue Type: Bug Components: Python Reporter: Wes McKinney Fix For: 0.6.0 {code} pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_python_file_write PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_python_file_read PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_bytes_reader PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_bytes_reader_non_bytes PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_bytes_reader_retains_parent_reference PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_buffer_bytes PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_buffer_memoryview PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_buffer_bytearray PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_buffer_numpy PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_buffer_memoryview_is_immutable PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_memory_output_stream PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_inmemory_write_after_closed PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_buffer_protocol_ref_counting PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_nativefile_write_memoryview PASSED pyarrow-test-2.7/lib/python2.7/site-packages/pyarrow/tests/test_io.py::test_mock_output_stream /Users/travis/build/apache/arrow/ci/travis_script_python.sh: line 81: 8186 Segmentation fault: 11 (core dumped) python -m pytest -vv -r sxX -s $PYARROW_PATH --parquet {code} -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (ARROW-1337) [Python] User reports pkg-config does not work properly in FindArrow.cmake for pyarrow
Wes McKinney created ARROW-1337: --- Summary: [Python] User reports pkg-config does not work properly in FindArrow.cmake for pyarrow Key: ARROW-1337 URL: https://issues.apache.org/jira/browse/ARROW-1337 Project: Apache Arrow Issue Type: Bug Components: Python Reporter: Wes McKinney Fix For: 0.6.0 {code} -- Checking for module 'arrow' -- Found arrow, version 0.5.0 -- Arrow ABI version: 0.0.0 -- Arrow SO version: 0 -- Found the Arrow core library: /gnu/store/h3cb0ynq76cmzs2vp2syqd42kkdh9paa-apache-arrow-0.5.0/lib/libarrow.so -- Found the Arrow Python library: /gnu/store/h3cb0ynq76cmzs2vp2syqd42kkdh9paa-apache-arrow-0.5.0/lib/libarrow_python.so CMake Error at cmake_modules/BuildUtils.cmake:88 (message): No static or shared library provided for arrow Call Stack (most recent call first): CMakeLists.txt:263 (ADD_THIRDPARTY_LIB) -- Configuring incomplete, errors occurred! {code} -- This message was sent by Atlassian JIRA (v6.4.14#64029)
Re: pyarrow versioning
Ah ok, makes sense -- thanks Wes! - Colin *Colin Nichols | Senior Software Engineer335 Madison Avenue, 16F | New York, NY, 10017+1 (646) 912 2018 | BAM.ai* On Mon, Aug 7, 2017 at 3:55 PM, Wes McKinneywrote: > hi Colin, > > Sorry about that. Yes, I pulled the 0.5.0 packages from PyPI (which is > an unofficial package, because the Arrow PMC has not voted on it) > because of the problems with jemalloc in ARROW-1282 -- it is not > possible to replace a broken package without also changing the version > number. This is a pretty exceptional case because the jemalloc > allocator was causing hung processes in some cases; this was something > we could disable at build time without making a new release. > > I recommend that you pin to a minor version, but not a patch version, so > > pyarrow==0.5.* > > Patch versions as a rule will not contain API changes. This conflicts > with the installation advice in http://arrow.apache.org/install/, so I > will update this at next opportunity. > > - Wes > > On Mon, Aug 7, 2017 at 3:40 PM, Colin Nichols wrote: > > Hi all, > > > > I noticed today that pyarrow==0.5.0 has disappeared from Pypi, replaced > by > > 0.5.0.post2. Just wanted to make sure that was intended. If so, is the > > expectation that users put e.g., pyarrow~=0.5.0 in their requirements > file > > as opposed to pyarrow==0.5.0? > > > > Thank you, > > Colin > > > > > > > > *Colin Nichols | Senior Software Engineer335 Madison Avenue, 16F | New > > York, NY, 10017+1 (646) 912 2018 | Narrativ* >
Re: pyarrow versioning
hi Colin, Sorry about that. Yes, I pulled the 0.5.0 packages from PyPI (which is an unofficial package, because the Arrow PMC has not voted on it) because of the problems with jemalloc in ARROW-1282 -- it is not possible to replace a broken package without also changing the version number. This is a pretty exceptional case because the jemalloc allocator was causing hung processes in some cases; this was something we could disable at build time without making a new release. I recommend that you pin to a minor version, but not a patch version, so pyarrow==0.5.* Patch versions as a rule will not contain API changes. This conflicts with the installation advice in http://arrow.apache.org/install/, so I will update this at next opportunity. - Wes On Mon, Aug 7, 2017 at 3:40 PM, Colin Nicholswrote: > Hi all, > > I noticed today that pyarrow==0.5.0 has disappeared from Pypi, replaced by > 0.5.0.post2. Just wanted to make sure that was intended. If so, is the > expectation that users put e.g., pyarrow~=0.5.0 in their requirements file > as opposed to pyarrow==0.5.0? > > Thank you, > Colin > > > > *Colin Nichols | Senior Software Engineer335 Madison Avenue, 16F | New > York, NY, 10017+1 (646) 912 2018 | Narrativ*
pyarrow versioning
Hi all, I noticed today that pyarrow==0.5.0 has disappeared from Pypi, replaced by 0.5.0.post2. Just wanted to make sure that was intended. If so, is the expectation that users put e.g., pyarrow~=0.5.0 in their requirements file as opposed to pyarrow==0.5.0? Thank you, Colin *Colin Nichols | Senior Software Engineer335 Madison Avenue, 16F | New York, NY, 10017+1 (646) 912 2018 | Narrativ*
Re: Arrow Plasma Object Store - IP clearance
Thanks! This is great! On Mon, Aug 7, 2017 at 11:30 AM Wes McKinneywrote: > Thanks to the Plasma developers for their code contribution and > efforts integrating it with the Arrow codebase! It's a powerful and > useful tool that will help the project grow. > > - Wes > > On Mon, Aug 7, 2017 at 2:24 PM, Philipp Moritz wrote: > > Great to hear! Thanks a lot to everybody involved with this for their > help. > > > > On Mon, Aug 7, 2017 at 11:19 AM, Julian Hyde wrote: > > > >> The vote for IP clearance of the Plasma Object Store on the Incubator > list > >> has passed[1]. > >> > >> We can now proceed with a release. > >> > >> Julian > >> > >> [1] https://s.apache.org/arrow-plasma-object-store-clearance-result > >> > >> > >> >
Re: Arrow Plasma Object Store - IP clearance
Thanks to the Plasma developers for their code contribution and efforts integrating it with the Arrow codebase! It's a powerful and useful tool that will help the project grow. - Wes On Mon, Aug 7, 2017 at 2:24 PM, Philipp Moritzwrote: > Great to hear! Thanks a lot to everybody involved with this for their help. > > On Mon, Aug 7, 2017 at 11:19 AM, Julian Hyde wrote: > >> The vote for IP clearance of the Plasma Object Store on the Incubator list >> has passed[1]. >> >> We can now proceed with a release. >> >> Julian >> >> [1] https://s.apache.org/arrow-plasma-object-store-clearance-result >> >> >>
Re: Arrow Plasma Object Store - IP clearance
Great to hear! Thanks a lot to everybody involved with this for their help. On Mon, Aug 7, 2017 at 11:19 AM, Julian Hydewrote: > The vote for IP clearance of the Plasma Object Store on the Incubator list > has passed[1]. > > We can now proceed with a release. > > Julian > > [1] https://s.apache.org/arrow-plasma-object-store-clearance-result > > >
Arrow Plasma Object Store - IP clearance
The vote for IP clearance of the Plasma Object Store on the Incubator list has passed[1]. We can now proceed with a release. Julian [1] https://s.apache.org/arrow-plasma-object-store-clearance-result
[jira] [Created] (ARROW-1336) [C++] Add arrow::schema factory function
Wes McKinney created ARROW-1336: --- Summary: [C++] Add arrow::schema factory function Key: ARROW-1336 URL: https://issues.apache.org/jira/browse/ARROW-1336 Project: Apache Arrow Issue Type: New Feature Components: C++ Reporter: Wes McKinney Fix For: 0.6.0 Because using {{std::make_shared}} with initializer lists is incompatible, it would be useful to have a factory function for making schemas from an initializer list of fields to make user syntax nicer -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Created] (ARROW-1335) [C++] PrimitiveArray::raw_values has inconsistent semantics re: offsets compared with subclasses
Wes McKinney created ARROW-1335: --- Summary: [C++] PrimitiveArray::raw_values has inconsistent semantics re: offsets compared with subclasses Key: ARROW-1335 URL: https://issues.apache.org/jira/browse/ARROW-1335 Project: Apache Arrow Issue Type: Bug Components: C++ Reporter: Wes McKinney Fix For: 0.6.0 {{NumericArray::raw_values}} accounts for offset, while {{PrimitiveArray::raw_values}} does not. This seems likely to lead to shooting one's self in the foot. It may be better to remove {{PrimitiveArray::raw_values}} altogether -- This message was sent by Atlassian JIRA (v6.4.14#64029)
Re: Arrow 0.6.0 release planning and timeline
hi all, It looks like we should be able to close out 0.6.0 issues today and cut a release candidate tomorrow if there are no objections. Please take a moment to complete any code reviews for patches that should go in. We have a number of Java library dependency upgrades pending, do these need to go in? https://github.com/apache/arrow/pull/929 https://github.com/apache/arrow/pull/873 There is another Java patch that needs to be merged if someone could review: https://github.com/apache/arrow/pull/898 There are a handful of C++ or Python patches pending, and I'll be opening a few more PRs today; any code reviews would be appreciated. Thanks Wes On Fri, Aug 4, 2017 at 3:57 PM, Siddharth Teotiawrote: > Reviewed https://github.com/apache/arrow/pull/915 for ARROW-1296 > > > On Fri, Aug 4, 2017 at 11:45 AM, Siddharth Teotia > wrote: > >> I will review it by EOD. >> >> On Fri, Aug 4, 2017 at 11:15 AM, Li Jin wrote: >> >>> On the Java side I have https://issues.apache.org/jira/browse/ARROW-1296, >>> which is small bug fix. >>> >>> If someone help review it would be great. Else if it doesn't get reviewed >>> by 0.6 rc cut, we can take it off 0.6 release. >>> >>> Li >>> >>> On Fri, Aug 4, 2017 at 2:02 PM, Wes McKinney wrote: >>> >>> > hi all, >>> > >>> > If there are no problems with the Plasma IP Clearance, I would like to >>> > cut a release candidate for 0.6.0 at the beginning of next week. There >>> > are a handful of issues pending on the Java and C++ side that I'll be >>> > working to complete over the next several days. Please keep an eye on >>> > the release page on JIRA: >>> > >>> > https://issues.apache.org/jira/projects/ARROW/versions/12341088 >>> > >>> > There are a number of outstanding Java patches; if you would like to >>> > include any of these in the 0.6.0 release, could someone review? >>> > >>> > Thanks, >>> > Wes >>> > >>> > On Tue, Aug 1, 2017 at 10:59 PM, Wes McKinney >>> wrote: >>> > > It seems that ARROW-1282 is causing some users problems. We have the >>> > > option of making a 0.5.1 release, but given how much work has reached >>> > > master (or is about to reach master) I would be in favor of >>> > > accelerating 0.6.0, cutting a release candidate within the next couple >>> > > of days. We could aim for another release within 2-3 weeks after >>> > > completing the Plasma IP clearance. >>> > > >>> > > Thoughts? >>> > > >>> > > On Tue, Aug 1, 2017 at 9:44 AM, Uwe L. Korn wrote: >>> > >> Hello, >>> > >> >>> > >> from my side we're mostly fine for a 0.6.0 release. Currently I'm >>> facing >>> > >> a problem with https://issues.apache.org/jira/browse/ARROW-1302 in >>> the >>> > >> 0.5.0 OSX wheels. We need to fix this before 0.6.0. Also I would >>> like to >>> > >> look a bit more into the jemalloc issues that came up with 0.5.0 to >>> get >>> > >> some of them solved in the next release. >>> > >> >>> > >> Uwe >>> > >> >>> > >> On Mon, Jul 31, 2017, at 04:55 PM, Wes McKinney wrote: >>> > >>> hi all, >>> > >>> >>> > >>> We're already 40 patches into the next Arrow version. I just created >>> > >>> https://issues.apache.org/jira/browse/ARROW-1297 as a tracking >>> issue >>> > >>> so that any blocking issues can be tracked as we push forward to >>> 0.6.0 >>> > >>> >>> > >>> You can track the status of the release here (accessible from the >>> > >>> "Projects" tab --> Releases in JIRA): >>> > >>> >>> > >>> https://issues.apache.org/jira/projects/ARROW/versions/12341088 >>> > >>> >>> > >>> We don't have any more data types slated for integration testing for >>> > >>> this release, but it might be nice to try to finish one or more of >>> > >>> them in the next week or two: >>> > >>> >>> > >>> - Fixed size binary >>> > >>> - Fixed size lists >>> > >>> - Decimal >>> > >>> - Union >>> > >>> >>> > >>> As far as timeline for 0.6.0, I would like to push for an RC the >>> week >>> > >>> of 8/14 at latest (assuming we are ready to ship the Plasma C++ >>> code), >>> > >>> reducing scope if needed. Any contributions of code, documentation, >>> or >>> > >>> JIRA prioritization would be much appreciated. >>> > >>> >>> > >>> Thanks, >>> > >>> Wes >>> > >>> >> >>