Product Feedback

  • Hot ideas
  • Top ideas
  • New ideas
  • My feedback
  1. Notebook export to JSON

    Ability to export notebook source (and possibly results) to JSON. Notebooks could then be easily parsed.

    4 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Notebooks  ·  Flag idea as inappropriate…  ·  Admin →
  2. Completed runs view should show more than 20 runs

    A common use case where 20 is a poor number is scheduled hourly jobs (minimum good default of 25, better default of 50 so that more than one full day may be seen).

    In addition, the user should be able to choose how many runs are shown per page.

    4 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Cluster management  ·  Flag idea as inappropriate…  ·  Admin →
  3. Fix the notebook filepath mouseover

    When the notebook name got moved to the left side of the notebook, the filepath mouseover still just shows up centered on the name.

    Then, when you click into the sidebar, the name of your notebook gets covered up.

    See attached screenshots.

    0 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Navigation UI  ·  Flag idea as inappropriate…  ·  Admin →
  4. Account setup: Default select time zone based on country/city, choose from map

    When setting up a new account, user is requested to enter country/city etc and then has to select a timezone.

    Clicking on the timezone box shows all available timezones and the user needs to choose one. Since there are many choices, it is hard to find the desired one.

    For most cases, the timezone should match the user's country/city so maybe default populate this field based on user country/city but also provide a world map pop-up with timezones/cities (similar to Linux setup)?

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Account management  ·  Flag idea as inappropriate…  ·  Admin →
  5. SparkNet was made available

    SparkNet runs Caffe neural net models in parallel. Vote for this to help machine learning run on Databricks!

    4 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  External libraries / applications  ·  Flag idea as inappropriate…  ·  Admin →
  6. Libraries could be loaded from a private Maven repo

    I'd like to use the "Maven Coordinates" feature for adding libraries to our Databricks cloud. The challenge is that this is for a private Maven repo. I see that the "Advanced" options when adding a library have a field for repository. That's a good start, but our repo is protected with basic http auth.

    Can you add support for specifying credentials for the repo?

    26 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    2 comments  ·  External libraries / applications  ·  Flag idea as inappropriate…  ·  Admin →
  7. Markdown cells from notebooks embedded with %run were rendered when a notebook is run as a job

    Currently, the Markdown of an embedded notebook only renders when the %run cell is run individually. If "Run All" is used or the notebook is run as a job, the Markdown from embedded notebooks is not displayed.

    This would be useful when building notebooks that are to act as dashboards. Often we want to generate the same visualization for various inputs which involves a parameterized notebook embedded in a notebook that defines the parameters.

    7 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Notebooks  ·  Flag idea as inappropriate…  ·  Admin →
  8. cluster names could be long-lived.

    For example, even beyond stops and starts (as in over the night when we shut down clusters or on weekends). Would be nice to always have a cluster name you could depend on. Might tie back to the other comments on ACL to a cluster.

    6 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    3 comments  ·  Cluster management  ·  Flag idea as inappropriate…  ·  Admin →
  9. Run Notebook During Cluster Creation

    To configure an elastic IP address, it would be great to run a notebook that assigns the IP address to the Spark Driver during cluster creation without having to manually run the notebook each time.

    4 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  10. Allow users to control Zoomdata instance

    Zoomdata does not need to be run 24/7 and we would like the ability to start and stop the instance running Zoomdata on demand.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  11. Enable GITHUB Integration at a basic subscription level (Please)

    I keep on coming back to one major thing about your product. We have a basic subscription. With that We do get a lot, but one thing would make our life easier is that you enable Github integration for basic subscriptions. Right now I can’t justify 10x the cost per month to upgrade to profession subscription just for Github integration. I can understand the price hike for enabling more and bigger clusters. Totally get that. But not for something simple like Github. Additional: Export is not sufficient enough because 1: DB export is proprietary and the human readable export is…

    7 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Other  ·  Flag idea as inappropriate…  ·  Admin →
  12. incompatibility: Python console is lacking sys.stdout.fileno, lets wget crash

    To reproduce the error include spacy library, then run:

    from spacy.en.download import main
    main()

    /home/ubuntu/databricks/python/local/lib/python2.7/site-packages/wget.pyc in getconsolewidth()

    142         winsize = array("H", [0] * 4)
    
    143 try:

    --> 144 ioctl(sys.stdout.fileno(), TIOCGWINSZ, winsize)

    145         except IOError:
    
    146 pass

    AttributeError: 'ConsoleBuffer' object has no attribute 'fileno'

    You can workaround the problem with following snippet:

    sys.stdout.fileno = lambda: 0

    This problem doesn't appear in IPython notebook, hence I assume the databricks notebook is lacking this attribute

    7 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Notebooks  ·  Flag idea as inappropriate…  ·  Admin →
  13. I could see list of my attached notebooks

    We have hundreds of notebooks. For easier workflow and navigation, it'd be cool if a user would be able to easily see a list of all his attached notebooks.

    This could be located on the left side menu, next to Recent.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Notebooks  ·  Flag idea as inappropriate…  ·  Admin →
  14. 1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Cluster management  ·  Flag idea as inappropriate…  ·  Admin →
  15. Allow external applications to call notebooks and receive datasets.

    Enable the ability for Dashboards and visualization front-ends to dynamically call (JDBC or Rest API) and pass in variables into SPARKR or other notebooks and receive datasets.

    6 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Notebooks  ·  Flag idea as inappropriate…  ·  Admin →
  16. 5 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  17. we could retrieve logs from messed up clusters that we had to reset

    When a DBC cluster is having problems, it can help to reset it. Sometimes there is pressure to do that quickly to get back up and running. After a reset, we should be able to see (and easily save an archive of) the old logs, so we can try to figure out what went wrong.
    Now, they seem to just disappear after a reset.

    6 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  18. display() can collect user input and make a new RDD

    We have some automatic classifier code. Not surprisingly, it is not perfect. If I make an RDD holding some input data and the resulting classification decision, I'd like to have a way to show those to the user so they can mark the decisions they know are right or wrong. That would then go into an active learning process where the classifier is retrained with the new data from each pass.

    What would be slick would be if display() or something like it could have a 'feedback' argument. When set, display()would do something like show columns of red X's and…

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Notebooks  ·  Flag idea as inappropriate…  ·  Admin →
  19. you would allow us to post more than 140 characters here as feedback

    ...or at least you let us know (by showing all the input boxes) that we would be able to describe things more fully before we just give up at the spread of the twitter meme.

    To clarify: now, the user interface just throws away all but the first 140 characters that we type or paste in, and gives no hints that we will also be asked for a category and a more complete description.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  20. mounted data showed up in the storage tab of the spark UI

    I mounted an S3 bucket and didn't realize the mount persisted after detaching my cluster, so was getting charged for extra SSD storage on AWS when I could've been unmounting the data after use everyday.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    2 comments  ·  Data import / export  ·  Flag idea as inappropriate…  ·  Admin →
  • Don't see your idea?

Feedback and Knowledge Base