Apache Arrow development container

• Includes all dependencies for Arrow development
• Builds are incremental, mirrored to local file system
• Resolves ARROW-2486

Get started

Install Docker

Acquire image

$ docker pull quiltdata/arrow

Populate host directory

Keep git repos and subsequent build products in a persistent local directory, /io.

$ mkdir -p io/arrow
$ git clone https://github.com/apache/arrow.git io/arrow
$ mkdir -p io/parquet-cpp
$ git clone https://github.com/apache/parquet-cpp.git io/parquet-cpp

Alternatively, if you wish to use existing git repos, you can nest them under /io.

Run container, mount /io as volume

$ docker run \
	--shm-size=2g \
	-v /LOCAL/PATH/TO/io:/io \
	-it quiltdata/arrow

Use container

Run scripts to build executables.

See also Arrow dev docs.

$ source script/env.sh
$ script/arrow-build.sh
$ script/parquet-build.sh
$ script/pyarrow-build.sh
# run tests
$ cd /io/arrow/python
$ py.test pyarrow

Build container

$ docker build -t USERNAME/arrow .