Looker 6.10 Release Notes


Userlevel 5
Badge

Anticipated Deployment Dates


Release Rollout Begins: April 7, 2019

Release Final & Download Available: April 18, 2019




Release Highlights


In addition to general tweaks and enhancements, this release comes with new and improved features in the following categories. Read on for more detail.



Preparing for Release


Please take notice of items marked with a ⚡ as they indicate changes to existing functionality and may require your attention. For more information see the Legacy Feature Updates and Features by Section below.



Experimental and Beta Features


The following new and improved experimental and beta features are marked with a ⚗:






Notable Features


Three (or four) New Visualizations


By popular request, we’ve added three new standard visualizations and a fourth Labs revamp of the table visualization. Use the new box plot to show distributions in value, use the waterfall to help understand the cumulative effect of positive and negative values, and highlight the most popular values with the new word cloud. The new table visualization has native subtotals and adjustable and freezable columns.





Features by Section


Dashboards, Visualizations, and Explore




  • New visualizations!



  • Reverse the colors. Added an option to reverse the order of a color palette when multiple series are present in a visualization.


Platform and Administration




  • Introduced Oauth for the Action Hub. New actions may be built with OAuth, allowing users to authenticate as themselves to take action in another application. Learn more.


  • New default password requirements. New and reset passwords will be subject to the following requirements: minimum length 10 characters, at least one uppercase and one lowercase letter, one numeric character, and one special character. Learn more.


  • New language support. Expanded current language support to include Spanish, Italian, Korean, Dutch, Portuguese, Brazilian Portuguese, and Russian. Learn more.


  • System Activity model and dashboard improvements

    • Fields used in non-Look tiles are now tracked in the model and dashboards.

    • Added descriptions to the remaining fields in the model.

    • Improved drill fields in the System Activity model and dashboards.




  • ⚗ Promoted API 3.1 from “experimental” to “beta”.

  • Added the ability to disable Gravatar user account profile pictures. Learn more.

  • Added Javascript events that track when embedded content begins and completes loading data. Learn more.


LookML and Development




  • The Content Validator now covers dashboard filters and listeners. Learn more.


Security



  • Addressed a vulnerability that allowed for access to other users’ scheduled deliveries.

  • Addressed a vulnerability related to the sanitization of SQL Runner downloads.

  • Newly created instance will default to having “Clickjacking Protection” enabled on the Embed Admin page. Existing instances will retain their selected values.


General Tweaks and Bug Fixes




  • Dialects

    • Fixed an issue where Looker generated invalid SQL for Dremio.

    • Fixed an issue where Looker generated incorrect SQL for relative date filters on Druid.




  • LookML and Development

    • Added a link to the “import credentials” project settings tab when the LookML Validator warns that a dependency lacks credentials.

    • Fixed an issue that resulted in “Document Not Found” messages when selecting “View Document” to see a document outside of the IDE.


    • ⚗ Custom fields no longer unnecessarily trigger Symmetric Aggregates.

    • Fixed an issue where developers could not see certain models connected to local git repositories.

    • Fixed an issue where the Content Validator could group errors by the wrong space.

    • Fixed an issue where the Content Validator did not flag errors produced by fields being excluded using fields: [] at the join level.

    • Fixed an issue where the LookML Validator failed to detect inaccessible native derived table fields.

    • Added validation to properly protect against duplicative model names issues.


    • ⚗ Fixed an issue where project import would include all view files in a project when it should just include the root with include: "/*.view".




  • Scheduling and Downloading

    • Improved mailer stability by retrying schedules after a timeout period.




34 replies

Userlevel 2

I’m intrigued about new experimental table viz, any other info about it?

Userlevel 5
Badge

@svickers they’re going to start out with subtotals and 🎉 adjustable and freezable columns 🎉 but the team has further refinements planned

Userlevel 5
Badge

Really impressed with how you guys are addressing your visualizations enhancement needs. This has been the only part of Looker that I’ve found lacking, but these improvements continue to be major!

Is there any plan to add French and Chinese language support?

Userlevel 5
Badge

Chinese is on the map, but it’s not currently of high priority, so no promises on if or when it’ll be integrated. French is already here! At least to the full extent all of our current localization efforts (nav, top-level menu items, etc. aren’t yet implemented)

Thanks for the information. Could you please confirm if there is any way in Looker by which in built messages such as ‘Loading Data…’ while loading the look can be translated to any available languages like French ?

Userlevel 5
Badge

Most strings on dashboards and visualizations should be localized, though I think the “Loading Data…” might be a release or so out. Let us know if anything’s missing for you and feel free to add any requests you might have to the feature requests category.

Thanks for the info.


Is there a fix for explore names that contain special characters, most notedly for us was the ‘&’?


Whenever we renamed an explore and then attempted to switch the name with the content validator, there were issues due to the ampersand being present in the old name.

Userlevel 3

I totally agree with @bens, the lack of depth on Viz options is the only thing I could be critical of Looker, so this is awesome news! We are looking forward to having Boxplots and an improved Table 🙂

Userlevel 5
Badge

I see many great improvement in system activity. Can we add filter to User Activity dashboard to exclude all users who are disabled?

Userlevel 7
Badge +1

FR: https://looker.atlassian.net/browse/ADMIN-604

Any updates on PDF’s with filters at the top?

Userlevel 7
Badge +1

Nothing at the moment, Dan— I know that there is some general dashboard redesigning in the pipeline, but I’m not certain whether that scope includes this feature. I’ll see if I can get someone from product to weigh in.

Userlevel 5
Badge

While it’s great to see these new visualizations, I feel like the waterfall and box plot are an iteration away from being ready for business users selecting them from an Explore (though maybe this is something with a few examples of how to use them properly from Looker, I’ll understand the intent more than I do right now). My concern is that it seems they both require using of Table Calcs or LookML, which to me goes against Looker’s raison d’être of no/low-code data dives for all.


For a Waterfall, for example, I’m not clear on how we’re intended to use it with one dimension and one measure. The most obvious use case for a Waterfall Chart is probably a visualization of income. For that, I would’ve thought the approach to accommodate would be to actually allow zero dimensions, and only measures. Then, to be able to specify whether each measure is an Up or Down contributor (i.e. flip the signs if needed), and the column headers could simply be the measure column names. With one dimension and one measure to accommodate this, pretty sure I’d have to write Waterfall-specific fields.


For a Box Plot, I understand the concept at least, but I still don’t think it’s overly usable to require writing between two and five measures just for it. The way I was hoping it would work, is to accept two dimensions and one measure and for the system to deduce the rest (the quartile locations, whisker extremities, etc.). Say your first dimension was Store Name, and your second was Order Number. Then, with a Total Sales measure, this could present a box per Store, with orders ‘bucketed’ into the appropriate regions.


Hopefully something in this ramble is of use to someone, but these are promising starts 🙂

Userlevel 1

For the Waterfall I have the transaction date as the dimension, and the sum of transactions as the measure.


When I first saw the topic of Waterfall charts it seemed to be about some kind of cost analysis, but my first thought was to show the balance of an account over time, and the Waterfall Chart option given recently covers this exactly without needing any calculated fields.


I can filter for client X with a load date of Y, and see how the value of those accounts has progressed, and get a good idea of how much is left on that account.


Even if you’re looking at costs A, B and C and incomings of X, Y and Z, I’d imagine you’d want all those in the same table/view and the dimension would be the transaction type or similar?

Userlevel 5
Badge

Interesting! Seems the use case I was envisioning is pretty different. For me, the thing I love about a waterfall chart is to take multi-variate realities (such as profit), and to deconstruct them into those raw components (such as + revenue - fixed costs - variable costs). That said, I totally neglected to think about that kind of use for Waterfalls, but I see the value completely!

Waterfall color options are buggy for me. Only let’s me specify a color if I have “Reverse colors” checked.

Are the online docs for the box plot available?

@Lewis_Osborne Here they are: https://docs.looker.com/exploring-data/visualizing-query-results/boxplot-options

Hi all, thanks so much for your feedback on the new visualizations! We love hearing from our customers on their use cases and where our features are working / not working for them. I do want to give a couple of updates around the waterfall and boxplot visualizations:


@bens I totally hear you on the 0 dimension use case for waterfall - we are going to be releasing an update soon that supports that configuration. We will also be pushing a couple of fixes to resolve bugs around the color picker - thanks for your patience as we roll these out!


For boxplot we did take an approach that is a bit different from other products given our in-database architecture. Since Looker does not consume or store the full dataset in memory, computing the min/max/quartiles based on the subset of rows available in the table could potentially result in inaccurate values. To ensure accurate results, we instead opted for the database to do the computation, which does require that measures or custom fields are needed to set it up.


Thanks again everyone for the feedback and helping us improve - keep it coming!


Cheers,

Arielle

I love the new table options - particularly the ability to autosize a column. Curious, will this new table support stacking measures verticlaly at some point? Tracking 4-5 measures in a vertical stack with time on the horizontal in a table is an incredibly common use case - think quarterly financial reports, P&L statements, etc.


Thanks for the great improvements!

Hi @divvyMatt, glad to hear you are enjoying the new table updates! Yes, the ability to transpose (flip the columns + rows) is in the works right now. Here’s a sneak peek 😀 (it’s not available quite yet, but will be in an upcoming release!)


Userlevel 5
Badge

Not to get too greedy here, but a collapse/expand all would be fantastic. It’s really cumbersome to use sub totals without this feature (especially since Looker is defaulting to expanded). @arielle thanks for your detailed response above! Looking forward to seeing those changes.

Userlevel 6
Badge +1

Christmas is coming early this year!

Hey @bens, collapse/expand all from the header bar is on our backlog, hoping to get that into a future update! In the meantime, you can actually change the default behavior in the viz config, here:


Reply