The Hue 4 user interface in detail

The Hue 4 user interface in detail

Aloha User Experience enthusiasts,

With the Hue 4 release we introduced a modern UI on top of our existing software to facilitate data discovery and analysis on premise and in the cloud.

The new UI organization

The new layout simplifies the interface and is now single page app, and this makes things snappier and unifies the apps together.

From top to bottom we have:

  • A completely redesigned top bar, with a quick action (big blue button), a global search and a notification area on the right
  • A collapsible hamburger menu that offers links to the various apps and a quick way to import data
  • An extended quick browse on the left
  • The main app area, where the fun is 🙂
  • A right Assistant panel for the current application. It’s now enabled for the editors, and in case of Hive for instance, it offers you a live help, a quick browse for the used tables in your query, and much more: if your Hue instance is connected to a SQL Optimizer service like Cloudera Navigator Optimizer, it can offer suggestions on your queries!

Various applications have been grouped into 4 main conceptual areas:

  • Editor: The goal of Hue’s Editor is to make data querying easy and productive. It focuses on SQL but also supports job submissions. It comes with an intelligent autocomplete, search & tagging of data and query assistance.
  • Browsers: Hue’s Browsers let you easily search, glance and perform actions on data or jobs in Cloud or on premise clusters.
  • Dashboard: Dashboards are an interactive way to explore your data quickly and easily. No programming is required and the analysis is done by drag & drops and clicks.
  • Scheduler: The application lets you build workflows and then schedule them to run regularly automatically. A monitoring interface shows the progress, logs and allow actions like pausing or stopping jobs.

The work on it started a couple of years ago and was incremental, and we applied a lot of the outcome of both design studies and customer feedback.

A few exotic locations helped on getting the inspiration:

Here it is, in all of its beauty 🙂


Quick search and browse

The new search bar is always accessible on the top of screen, and it offers a document search and metadata search too if Hue is configured to access a metadata server like Cloudera Navigator.

The improved quick browse on the left side of the screen now offers many more data sources and it is now enabled for Hive and Impala (like it was before) but also for HDFS, S3, HBase, Solr collections and Hue documents.


Default action / landing page

In Hue 4, every user is able to set their own preferred main action (the big blue button) and landing page. It just takes one click on the star next to the application name, like in this case next to Hive

The next time you will login, you will land directly into the Hive editor and a new query is always one click away.


Backward compatible

The older Hue 3 UI is still there and it’s easily reachable just by clicking on ‘Switch to Hue 3/4’ from the user menu.

Administrators can also decide to enable/disable the new UI at a global level on the hue.ini or CM safety valve

# Choose whether to enable the new Hue 4 interface.

If you look at your browser’s address bar, you will notice that all the URLs with the /hue prefix point to Hue 4. It is possible to just remove the prefix and land on the Hue 3 version of the page, e.g. /hue/editor (Hue 4) → /editor (Hue 3)

Now go try it on or download it if you haven’t done it already!



  1. Kevin LEFEVRE 1 year ago

    Hello, does HUE 4.0 compatible with Hortonworks HDP 2.6 ?

    • Kevin LEFEVRE 1 year ago


    • Author
      Hue Team 1 year ago

      Yes, Hue is compatible with any Hadoop distribution (the standard API did not change much).

  2. Ann 1 year ago

    Hi, we are in the process of switching to a new cluster with Hue 4 & Cloudera 5.7. In our old cluster (with Hue 3), the Oozie UI had an Email action as one of the icons across the top of the screen. It’s missing in the new cluster with Hue 4 and our Hadoop Admins can’t determine how to add it. Switching to the Hue 3 UI on the new cluster doesn’t make it appear. Any ideas on how to get the Email Action to to appear in Oozie on the new cluster? We like the new “email-on-failure” option at the end of a workflow but still need to be able to add email actions in the middle of a workflow or for success conditions too.

  3. Radhakrishnan Rk 1 year ago

    Can I disable hue 3 and keep only hue 4 globally

  4. Sandeep 1 year ago

    I cant find “Actions” tab for a workflow, but this is available in Hue 3 which is a tabular view and easy to get information than a graph. I can see only “Graph” view in Hue 4. Am i missing something or is it removed in Hue 4.

    • Author
      Hue Team 1 year ago

      Don’t you have ‘Task’ in the Job Browser page of a workflow?

  5. Georg 1 year ago


    I have one question and a few pieces of criticism on the new UI that I hope can be improved:
    How do I see my existing Oozie workflows & coordinators?
    In Hue 3 it was easy: Workflows -> …
    In Hue 4 I can’t find them at all.

    What’s an easy way to switch between Hive & Impala? In Hue 3 it was one click and one vertical mouse movement.
    In Hue 4 it’s: Try to hit the small arrow next to “Query” -> vertical mouse movement to “Editors” -> Horizontal mouse movement (trying to stay in line) -> potentially another vertical mouse movement.

    In Hue 3 it was super easy to switch between apps because all of them were in the top bar, now I need to use the Hamburger menu (extra movement & click).

    This was already a problem in Hue 3 but the Oozie Action icons are pretty much not self-explanatory. It’d be great to have a text explaining what they are.

    The “Documents” stuff in Oozie is super confusing. I want to see all my actions, I don’t even know what “Documents” are supposed to be….yet another click.

    I agree that the whole thing looks much nicer but in my opinion it’s harder to use (the SQL editor etc. are great, I mean the navigation in Hue itself)

    • Author
      Hue Team 1 year ago

      Thanks for the great and valid feedback!

      The whole idea of the new UI is to avoid any page change and context switch. It needs a bit of time to re-adjust and some adjustments here and there are still needed.

      #1 How do I see my existing Oozie workflows & coordinators?
      The top right Oozie menu points to the list of workflows in the home page. Is it too hidden?
      A next step is to automatically filter by the document type in the left doc assist, so that you would get all your workflow if you are in the Workflow Editor, all Hive queries if in Query Editor etc…

      #2 What’s an easy way to switch between Hive & Impala?
      In practice we found either people user one of the other, but rarely switch between. What is your use case to need to switch often? (curious about it)
      In Hue 3 it all started to look very complicated and many end users preferred to just see one Editor, Dashboard etc as they have no notion of Hadoop.

      #3 This was already a problem in Hue 3 but the Oozie Action icons are pretty much not self-explanatory.
      Indeed, and this is why we are switching to documents, and there is a prototype with only one automatic action, where the user does not need to know anything. The ‘documents’ also needs to be tweaked. The good news is that the Editor and Dashboard are getting pretty good (5000 improvement commits this year), and next on the list is the Oozie scheduler, so expect much more polishment there next year.

      • Georg 1 year ago

        Thanks for the answer.

        #1: I’ve just looked and with your help I was able to find them. Going to the “Scheduler” App which opens a new Workflow and then clicking a button to see your existing ones seems very counter-intuitive to me.

        #2: It might just be me but I switch between the two all the time. One is good for interactive stuff and the other for ETL like queries and longer running things. Some things are not supported in one but in the other or I do a metadata refresh in Impala after doing something in Hive. The same happens when I give trainings and ask people to compare the two etc.

        #3: This might be just me but I totally don’t understand what the documents are about like…at all (and I’ve used Hue since version 1). All I need is to place actions and for that I don’t want to guess what the icons are meant to be.

        I realize that some of this is due to personal preference and I’m sure you made all these changes in good faith. All I’m trying is to offer a different view on things. I’ve just finished a three day training and 10 out of 13 people (all total newcomers to Hadoop) switched back to the Hue 3 interface at some point during the training.

        • Author
          Hue Team 1 year ago

          More focus on the Oozie part will come in 2018, currently most of the effort is in getting the SQL Editor rock solid and the flow from there smoother. Most of the regular users often just stick to the Impala Editor and don’t go too far from there. What are the goals of the 13 new people? (do SQL queries or more ETL/Job Scheduling etc..)

  6. Tony 1 year ago

    Where did the metastore manager go?

    • Author
      Hue Team 1 year ago

      It’s available from the left menu “Browsers > Tables” or by right clicking on any table in the left table listing and “Opening in Browser”.

  7. Eric 1 year ago

    If you use Hive mostly, having the Query button remember your last used editor, or going to the source selected in the Quick Browse panel, probably would be much more useful than having it always to go the Impala editor. It’s only an extra click (and some moderately finessed mouse movement), but still.

    • Author
      Hue Team 1 year ago

      By hovering over the editor name, you will see a star icon appear. If you press the star, you will favorite the editor and it will be used as default.

  8. Amanda Holman 12 months ago

    How do you turn on the grey box where the log is going in real-time when you are running query? Mine used to be turned on then I accidentally clicked the ‘x’ in the upper corner and it has disappeared. I don’t see where you can show that display anymore.

    • Author
      Hue Team 12 months ago

      When the query is running, you should see a little blue triangle toggle icon just on the top left of the progress bar.

  9. Sips 10 months ago

    The import button as suggested in one of Hue demos is missing. I want to import a JSON workflow. My cluster is a new Azure HDInsight-R cluster. Is the something I am missing?

    • Author
      Hue Team 10 months ago

      The versions of Hue distributed by other companies and in the cloud (ie Microsoft’s Azure, AWS, etc) might differ from what we describe on the blog for architectural or commercial reasons, so it might be that the import functionality is not available.

  10. kant 9 months ago

    I have installed hue 4 with HDP 2.6 , i am abale to query all table getting good result. when i am trying to look in to graph i am not able to see the graph for the few tables. my all table s are extrenal hive tables.
    in few table like examples table and all dispalying graph very nice , but not for my datasets.
    kindly help ..?

    • Author
      Hue Team 9 months ago

      Which graph exactly?

  11. Dattaram 9 months ago

    Can I disable Details from “show details” for views created?

    • Author
      Hue Team 9 months ago

      Currently it’s not possible to hide part of the details. What would you like to hide and what’s the use case? Thanks!

  12. Apoorv 8 months ago

    I am unable to create a new notebook, which I used to do in Hue 3:


    However, with Hue 4, I am unable to find this feature.

    Could you please let me know how to create a new notebook?

  13. Andrew 7 months ago

    how do you hide/unhide the quick browse and assist columns??

    • Author
      Hue Team 7 months ago

      On the top left, there is a little toggle showing-up on hover

  14. Fred 6 months ago

    I have a lot of Pig Scripts, created with pig editor in Hue 3.
    After Hue 4 migration (CDH 5.15.1), all my scripts are seen as “pig design” and not as “script pig”.
    And when i click on a “pig design” doc (or right click and open document), nothing !
    Impossible to have the content of my scripts.

    And Hue 3 pig editor doesn’t work anymore. Blanck screen.

    How can I open “pig design” document in Hue 4 ?


    • Author
      Hue Team 5 months ago

      Yes, the Pig Editor was moved to the common Editor and should have been converted to the Hue 4 format:

      Do you have a Javascript console error? (right click in the Chrome tab)
      Could you export one of your Pig script via the homepage and share it?

  15. federico 4 months ago

    I’m running CDH 5.16 with Hue 4.x version on Aws Ec2 (32 Gb ram, 44 cores,200 gb Disk).
    We managed to launch jobs(Hive/2, Shell), generate workflows and scheduler.

    On scheduler we meet a problem :
    New Scheduler-> give a name-> choose workflow -> set how often (here is the issue)

    1)we set Timezone on ‘Europe/Rome’, ‘from ‘and ‘to’ date

    the hour for “from Date” and “To Date” have a time difference with our timezone.
    if there is written “from 11/19/2018 05:10” we found that 5:10 is not our 5:10 pm in Europe/Rome timezone
    but it is a guess, because for every hour that is set into “From/ To” date, we are not able to understand when the scheduling starts.

    How can we set the hour HH:mm in order to be executed at HH:mm Europe/Rome Timezone ?

    Why did you take out the frequency panel? if I want to set a frequency (schedule this job every xx minute, I can only do it by command 0/5****

  16. Fred 1 month ago

    using PIG and Hue4.
    All hcat functions doesn’t work anymore in HU4 (hcatloader, …..etc) .
    Why this choice ?

    • Author
      Hue Team 1 month ago

      Do you mean they do not work out of the box (are not imported automatically)? What is the error?

  17. pete 2 weeks ago

    hello, is it possible to manipulate for example the column names and? I’m currently using Hue hosted in emr hive using AWS glue as metastore. Tried as well to add description but I’m getting this error.

    “traceback”: [ [ “/usr/lib/hue/build/env/lib/python2.7/site-packages/Django-1.11-py2.7.egg/django/core/handlers/”, 41, “inner”, “response = get_response(request)” ], [ “/usr/lib/hue/build/env/lib/python2.7/site-packages/Django-1.11-py2.7.egg/django/core/handlers/”, 249, “_legacy_get_response”, “response = self._get_response(request)” ], [ “/usr/lib/hue/build/env/lib/python2.7/site-packages/Django-1.11-py2.7.egg/django/core/handlers/”, 187, “_get_response”, “response = self.process_exception_by_middleware(e, request)” ], [ “/usr/lib/hue/build/env/lib/python2.7/site-packages/Django-1.11-py2.7.egg/django/core/handlers/”, 185, “_get_response”, “response = wrapped_callback(request, *callback_args, **callback_kwargs)” ], [ “/usr/lib/hue/build/env/lib/python2.7/site-packages/Django-1.11-py2.7.egg/django/utils/”, 185, “inner”, “return func(*args, **kwargs)” ], [ “/usr/lib/hue/desktop/core/src/desktop/”, 97, “hue”, “‘cluster’: clusters[0][‘type’] if clusters else None” ], [ “/usr/lib/hue/desktop/core/src/desktop/lib/”, 223, “render”, “return render_json(data, request.GET.get(\”callback\”), status=status)” ], [ “/usr/lib/hue/desktop/core/src/desktop/lib/”, 298, “render_json”, “json = encode_json(data, indent)” ], [ “/usr/lib/hue/desktop/core/src/desktop/lib/”, 270, “encode_json”, “return json.dumps(data, indent=indent, cls=Encoder)” ], [ “/usr/lib64/python2.7/json/”, 251, “dumps”, “sort_keys=sort_keys, **kw).encode(obj)” ], [ “/usr/lib64/python2.7/json/”, 209, “encode”, “chunks = list(chunks)” ], [ “/usr/lib64/python2.7/json/”, 434, “_iterencode”, “for chunk in _iterencode_dict(o, _current_indent_level):” ], [ “/usr/lib64/python2.7/json/”, 408, “_iterencode_dict”, “for chunk in chunks:” ], [ “/usr/lib64/python2.7/json/”, 408, “_iterencode_dict”, “for chunk in chunks:” ], [ “/usr/lib64/python2.7/json/”, 442, “_iterencode”, “o = _default(o)” ], [ “/usr/lib/hue/desktop/core/src/desktop/lib/”, 76, “default”, “return json.JSONEncoder.default(self, o)” ], [ “/usr/lib64/python2.7/json/”, 184, “default”, “raise TypeError(repr(o) + \” is not JSON serializable\”)” ] ] }

Leave a reply

Your email address will not be published. Required fields are marked *


This site uses Akismet to reduce spam. Learn how your comment data is processed.