Source Material for using Python and Hadoop together.
This is a tutorial for using Ibis and PySpark to interact with data stored in Hadoop, particularly files in HDFS and Impala Table.
You will need access to a Hadoop cluster (or a VM/Docker image), a python interpreter with the packages listed in requirements.txt installed, and Spark 1.6.1 installed as described in the 'Linking with Spark' section of the docs .
We will use hourly Wikipedia page view statistics that have been corrected: link. More documentation on the data source can be found on the [wikitech wiki page for the dataset.] (https://wikitech.wikimedia.org/wiki/Analytics/Data/Pageviews)
To download the data locally, run
mkdir pageviews-gz
python grab_data.pyfrom the root of this directory.
To then create the required tables,