As noted in BEAM-8368 it seems like pinning to 0.13.0 is a workaround for now. It would be useful to know what pyarrow versions have this problem, is it just 0.14, or 0.15 as well? Could someone with macOS 10.15 try pip installing pyarrow==0.15.0 and see if they can import it? We currently have a <0.15 bound, but maybe relaxing that will alleviate this issue.
If this is still a problem with pyarrow 0.15 we should file a jira with that project (and maybe restrict our requirement to <0.14 for on macos?). Brian On Thu, Oct 10, 2019 at 12:26 PM Kyle Weaver <[email protected]> wrote: > Apparently this is an issue with our arrow import. There is a JIRA for > this now: https://issues.apache.org/jira/browse/BEAM-8368 > > On Thu, Oct 10, 2019 at 12:11 PM Robert Bradshaw <[email protected]> > wrote: > >> Looks like an issue with the protobuf library. Do you know what >> version of protobuf you're using? (E.g. by running pip freeze.) >> >> I don't have Catalina to test this on, but it'd be useful if you could >> winnow this down to the import that fails. >> >> On Thu, Oct 10, 2019 at 8:15 AM Kamil Wasilewski >> <[email protected]> wrote: >> > >> > Hi all, >> > >> > I've recently updated my macOS to 10.15 Catalina. Since then, I have >> the following error when I try to import apache_beam package (both in >> python 2.7 and 3.x): >> > >> > >>> import apache_beam >> > [libprotobuf ERROR google/protobuf/descriptor_database.cc:58] File >> already exists in database: >> > [libprotobuf FATAL google/protobuf/descriptor.cc:1370] CHECK failed: >> GeneratedDatabase()->Add(encoded_file_descriptor, size): >> > libc++abi.dylib: terminating with uncaught exception of type >> google::protobuf::FatalException: CHECK failed: >> GeneratedDatabase()->Add(encoded_file_descriptor, size): >> > >> > [1] 43669 abort python >> > >> > Does anyone has the same problem? I saw that someone else had the same >> error at the beam-python slack channel, so I guess this problem is not >> limited to my workstation. >> > >> > Thanks, >> > Kamil >> >
