How to install pyarrow in Amazon Linux 2 m6g instance

Viewed 1225

I'm trying to install pyarrow in Amazon Linux 2 m6g instance which have ARM architecture. I have installed several dependencies and now I'm stuck on this error when I run pip3.8 install pyarrow:

-- Found the Arrow core static library: /usr/lib64/libarrow.a
    -- Could NOT find ArrowPython (missing: ArrowPython_DIR)
    -- Checking for module 'arrow-python'
    --   No package 'arrow-python' found
    CMake Error at /usr/local/share/cmake-3.18/Modules/FindPackageHandleStandardArgs.cmake:165 (message):
      Could NOT find ArrowPython (missing: ARROW_PYTHON_INCLUDE_DIR
      ARROW_PYTHON_LIB_DIR) (found version "2.0.0")
    Call Stack (most recent call first):
      /usr/local/share/cmake-3.18/Modules/FindPackageHandleStandardArgs.cmake:458 (_FPHSA_FAILURE_MESSAGE)
      cmake_modules/FindArrowPython.cmake:76 (find_package_handle_standard_args)
      CMakeLists.txt:215 (find_package)

It says that I need to install the package ArrowPython. When I run sudo yum install arrow-python-libs to install this package I get this error:

Error: Package: openblas-threads-0.3.3-2.el7.aarch64 (epel)
           Requires: libgfortran.so.3()(64bit)
Error: Package: openblas-threads-0.3.3-2.el7.aarch64 (epel)
           Requires: libgfortran.so.3(GFORTRAN_1.0)(64bit)
Error: Package: openblas-serial-0.3.3-2.el7.aarch64 (epel)
           Requires: libgfortran.so.3(GFORTRAN_1.0)(64bit)
Error: Package: python36-numpy-1.12.1-1.el7.aarch64 (epel)
           Requires: libpython3.6m.so.1.0()(64bit)
Error: Package: python36-numpy-1.12.1-1.el7.aarch64 (epel)
........
Error: Package: openblas-serial-0.3.3-2.el7.aarch64 (epel)
           Requires: libgfortran.so.3()(64bit)
 You could try using --skip-broken to work around the problem
 You could try running: rpm -Va --nofiles --nodigest

I have tried also with other pyarrow versions but I'm getting the same error.

How can I solve this?

1 Answers

I couldn't install via conda. Or the conda one had Import errors when I try import pyarrow. But got it working. With conda installed, via:

pip install pyarrow=2

Related