patelnilay / dremio_client

Un-official Python client for Dremio

Dremio client

The un-official python client for Dremio's REST API. This enables both administrators and data scientists to get the most out of Dremio in Python

• Documentation: https://dremio-client.readthedocs.io.
• Github: https://github.com/rymurr/dremio_client
• PyPI: https://pypi.python.org/pypi/dremio_client
• Free software: Apache Software License 2.0

Features

• Cross platform support
• All Pythons between 2.7 - 3.7 supported
• Full support for Dremio's REST API
• Optional Support for Dremio's ODBC or experimental Arrow Flight capabilities

• Rich config file support via confuse yaml config library. Simple to create a client (config stored in a yaml file)

  from dremio_client import init
  client = init() # initialise connectivity to Dremio via config file
  catalog = client.data # fetch catalog
  vds = catalog.space.vds.get() # fetch a specific dataset
  df = vds.query() # query the first 1000 rows of the dataset and return as a DataFrame
  pds = catalog.source.pds.get() # fetch a physical dataset
  pds.metadata_refresh() # refresh metadata on that dataset

• CLI interface for integration with scripts

  $ dremio_client query --sql 'select * from sys.options'
  {'results':results}
  $ dremio_client refresh-metadata --table 'my.vds.name'
  {'result':'ok'}

• Catalog autocompletion in Jupyter Notebooks

Status

This is still alpha software and is relatively incomplete. Contributions in the form of Github Issues or Pull requests are welcome. See CONTRIBUTING

TODO

• put, delete and post
• add reflections into catalog
• test for larger spaces and ensure it doesnt fetch everytihgn if we hit directly
• search
• docs
• logging
• handle exceptions correctly in cli
• osx & appveyor
• testing