February 2018

Releases are staged. Your Azure Databricks account may not be updated until a week after the initial release date.

New line chart supports time-series data

Feb 27 - Mar 6, 2018: Version 2.66

A new line chart fully supports time-series data and resolves limitations with our old line chart option. The old line chart is deprecated, and we recommend that users migrate any visualizations that use the old line chart to the new one.

Line Chart

See Migrate Deprecated Line Charts for more information.

More visualization improvements

Feb 27 - Mar 6, 2018: Version 2.66

You can now sort columns in table output and use more than 10 legend items in a chart.

Delete job runs using Job API

Feb 27 - Mar 6, 2018: Version 2.66

You can now use the Job API to delete job runs, using the new jobs/runs/delete endpoint.

See Runs Delete for more information.

KaTeX math rendering library updated

Feb 27 - Mar 6, 2018: Version 2.66

The version of KatTeX that Azure Databricks uses for math equation rendering was updated from 0.5.1 to 0.9.0-beta1.

This update introduces changes that can break expressions that were written in 0.5.1:

  • \xLongequal is now \xlongequal (#997)
  • [text]color HTML colors must be well-formed. (#827)
  • \llap and \rlap now render contents in math mode. Use \mathllap (new) and \mathrlap (new) to provide the previous behavior.
  • \color and \textcolor now behave as they do in LaTeX (#619)

See the KaTeX release notes for more information.

Databricks CLI: 0.5.0 release

February 27, 2018: databricks-cli 0.5.0

Databricks CLI now supports commands that target the Libraries API.

The CLI also now supports multiple connection profiles. Connection profiles can be used to configure the CLI to talk to multiple Azure Databricks deployments.

See Databricks CLI for more information.

DBUtils API library

Feb 13-20, 2018: Version 2.65

Azure Databricks provides a variety of utility APIs that let you work easily with DBFS, notebook workflows, and widgets. The dbutils-api library accelerates application development by allowing you to locally compile and run unit tests against these utility APIs before deploying your application to an Azure Databricks cluster.

See Databricks Utilities API library for more information.

Filter for your jobs only

Feb 13-20, 2018: Version 2.65

New filters on the Jobs list let you display only the jobs you own and only the jobs you have access to.

Job Filters

See Jobs for more information.

Spark-submit from the Create Job page

Feb 13-20, 2018: Version 2.65

Now you can configure spark-submit parameters from the Create Job page, as well as through the REST API or CLI.

Spark-submit

See Jobs for more information.

Select Python 3 from the Create Cluster page

Feb 13-20, 2018: Version 2.65

Now you can specify Python version 2 or 3 from the new Python version drop-down when you create a cluster. If you don’t make a selection, Python 2 is the default. You can also, as before, create Python 3 clusters using the REST API.

Python Version

See Python version for more information.

Workspace UI improvements

Feb 13-20, 2018: Version 2.65

We have added the ability to sort files by type (folders, notebooks, libraries) in the Workspace file browser, and the home folder always appears at the top of the Users list.

Workspace Sort

Autocomplete for SQL commands and database names

Feb 13-20, 2018: Version 2.65

SQL cells in notebooks now provide autocompletion of SQL commands and database names.

Serverless pools now support R

Feb 1-8, 2018: Version 2.64

You can now use R in serverless pools.

XGBoost available as a Spark Package

Feb 1-8, 2018: Version 2.64

XGBoost’s Spark integration library can now be installed on Azure Databricks as a Spark Package from the Library UI or the REST API. Previously, XGBoost required installation from source via init scripts and thus a longer cluster start-up time. See XGBoost for more information.

Table access control for SQL and Python (Beta)

Feb 1-8, 2018: Version 2.64

Last year, we introduced data object access control for SQL users. Today we are excited to announce the public beta release of table access control (table ACLs) for both SQL and Python users. With table access control, you can restrict access to securable objects like tables, databases, views, or functions. You can also provide fine-grained access control (to rows and columns matching specific conditions, for example) by setting permissions on derived views containing arbitrary queries.

Note

  • This feature is in public beta
  • This feature requires Databricks Runtime 3.5+.

See Table Access Control for more information.