Looker 7.10 Release Notes


Userlevel 5
Badge

As Looker is not officially posting updates here anymore then I thought to still share it as this is one of the best places for customers to give feedback about new updates.


I have listed first the official highlights provided by Looker and then added unofficial highlights by roles, which is something I share usually to our own users.


Let me/us know what do you think about those updates?




Expected Rollout Start: June 14th, 2020


Expected Final Release and Download Available: June 25th, 2020






Highlights provided by Looker:


Looker API CORS

A new experimental Labs feature that permits browser-based applications to directly call the Looker API without needing a proxy server for the API calls.


💰 Dashboards (Beta): Filter Themes

Introducing the ability to customize embed filter colors.


Dashboards (Beta): Enhancements

Introducing new menu options on dashboard (beta) tiles and the ability to retrieve dashboard LookML from dashboards (beta).


Database Connection Pooling

A new beta Labs feature that can improve query performance and support more concurrent users for supported database dialects.




Unofficial highlight by user groups:


Regular user (Browse menu):



  • Users can create a board from the Browse menu.


Power-user (Explore):



  • The inability to edit multiple dashboard filters simultaneously has been addressed.

  • The Calendar Heatmap custom visualization available on the Looker Marketplace includes the following improvements: new styles, new formatting options, the ability to drill, and table calculation support.


  • pngwing.com The Looker Marketplace now offers a Community Mobility Data Block that’s included within the COVID-19 Data Block.

  • The ACS Demographic Data Block has been added to the Looker Marketplace.


Developer (Develop):



  • A new Liquid parser is now used during LookML validation. This new parser will raise new Liquid warnings for syntax issues that were previously ignored. The warnings will indicate the line of LookML code causing the warning. https://help.looker.com/hc/en-us/articles/360049663734

  • Python SDK support for Python 3.6 is now available.

  • Refinements can now be used in Aggregate Table LookML for dashboards and Explores.

  • The extends LookML parameter can now be implemented within refinements.


Admin (Admin):



  • A new beta Labs feature, Database Connection Pooling, is now available. This feature optimizes database connection re-use via the JDBC driver, which allows for the maintenance of a pool of connections when databases with supported dialects are accessed.

  • A new experimental Labs feature, Looker API CORS , is now available. This feature allows browser-based web applications to call the Looker API across domain boundaries using the Cross Origin Resource Sharing (CORS) protocol. Each web application must be registered with the Looker instance and enabled by a Looker admin. Disabling this feature immediately blocks all cross-domain web requests to the Looker API.

  • A user with the create_alerts permission no longer sees the Admin panel without any menu options.


19 replies

Userlevel 5
Badge

Just overall comment from my side about release notes then I think it’s a bit of cheating to announce Dashboard (beta) new features as main highlights for the things that we already have for the original dashboards.


Some examples:


7.10: Get LookML

7.4: Download

7.2: Alerts

and so on.

Userlevel 6
Badge +1

I am quite surprised, concerned, and even unhappy about the releases lately.


In my mind Enhancements shouldn’t be highlights, after all, you’re just improving something still in Beta phase.


To me the releases miss a lot of bug fixes and small features that have been requested for months and even years

Userlevel 5
Badge

Unfortunately must agree here. I truly believe that there are many people in the company who are doing an amazing job of pushing the company forward but looks like something is just not working out.


What I feel is that company is going through a phase of reinventing themselves. Looker is shifting from the original path (exploring the data, lookml dashboards, etc) to more of a micro apps generator(white-label, embedding, colour changes).


Also to illustrate expectations vs reality I must include slides from 2019 roadmap. 😅



Userlevel 3

Thanks for the really thoughtful feedback, @er1k and @Dawid_Nawrot. It’s always helpful to hear how what we’re delivering is being received. On the point specifically about how we deliver release notes and what counts as a “highlight,” we moved release notes and change logs to our docs site in an attempt to make them more accessible and centralized. The goal wasn’t to get them off Discourse, and I know our community manager is thinking about ways to post things here as well while still keeping things in sync. So hopefully you’ll see some improvements on that front soon.


To the point about what we choose to “Highlight,” that word may be introducing more confusion than we realized. The goal of pulling those items out of the fuller changelogs is really just to give them more visibility since we expect they’re the changes that will affect the most users, and since many readers don’t delve into the full changelogs, they’re the things we want to make sure they’re aware of.


We’ll continue to think about ways to better target different announcements to different types of users, and I think Erik’s really great breakdown of features by user type is something we should look at closely to see if we could replicate it.


Finally, one thing that’s not highlighted in the release notes, but that I’m really excited about, is our move to a new system for collecting feature requests from users. As of 7.10, you should be able to see a new link in the Help dropdown called “Product Idea?” that will lead you to a portal where you can suggest ideas and vote up those of other users. This more structured data collection should help us identify the things that are affecting the most users.


You can read more about the new process here: How Your Feedback Makes Looker Better

Userlevel 6
Badge +1

Thank you @Mintz for your response.


You’re right. The word highlight might bring confusion but you say that you expect those changes that are considered highlights to affect most users. Personally I can’t imagine how fixes to Beta feature can be considered highlights in the scope of the whole product. How do you know what affects most users? I understand that Community is just a small part of your requests but we have no visibility of any product roadmap to make us understand the bigger picture.


Don’t get me wrong, I think Beta Dashboards will be an important part of Looker one day but @er1k has already mentioned that announcing fixes to make Beta Dashboards on par with the current ones do not feel like highlights.


I’ve heard about the Product Idea and I’m looking forward to seeing it in action! 🙂


I think at this point, we just need some clarity as to whether things we’ve been concerned about for the last few years are going to be fixed/improved.

Userlevel 5
Badge

Thanks, @Mintz from my side also. Always good to see that the community is being heard. 🙂 Hoping great results from “Product Idea?”. 🤞


Overall, I know that the company has had an interesting year but now I would love to see some proper progress with core functionalities. Like some basic viz functionalities (annotations, histograms, actual pivoting) or finally some highly hoped improvements/changes to lookML dashboards. So we wouldn’t need to worry about what is added or not added in the “highlights”. 🙏

Userlevel 1

Not mentioned here but I’m very happy to see the improved groups page. It’s a small improvement but much needed for customers like us with 100+ groups. Hoping we’ll see more improvements targeted to embedded uses cases.


Agree with the comments about dashboards beta. But we are looking forward to the full release and the upgrades for LookML dashboards(?)!

For years I’ve been looking forward to each months release notes and must say I enjoy the new format. Previously I’d read everything from the change log and highlight the big stuff for my team, and now it’s mostly done for us. Erik’s categories would be another welcomed improvement.


The dashboard-beta feature parity releases haven’t been so exciting for us either, but I trust there’s some bigger drops in the works. Personally, I’d like to see some more fundamental improvements like content management. The product idea portal sounds like a great idea.


@Mintz I appreciate your response and the whole Looker team ✊

Userlevel 7
Badge +1

I also love Erik’s user categories! In our haste to bring the new format to life, we stopped posting the notes here— But rest assured they are coming back shortly, and I’d like to incorporate Erik’s role based breakdowns.

Userlevel 5
Badge

Maybe you can switch from 7.8 to 7.10 in the community top banner also? 🙂

I’m not a fan of the new “Release Notes”. It highlights upcoming features but totally neglects major changes that have the potential to break a clients entire project. These are now reverted to 2 or 3 sentences in the change-log, with no examples or references to explain how new logic may work or what changes need to be made. I would go as far as saying it’s shoddy.

The notes/examples around how this works need fleshed out.


A new Liquid parser is now used during LookML validation. This new parser will raise new Liquid warnings for syntax issues that were previously ignored. The warnings will indicate the line of LookML code causing the warning. https://help.looker.com/hc/en-us/articles/360049663734


This new liquid parser has completely broken our project, as we used “_filters” to build urls and links across dashboards/looks that may not be joined at the model level. Clients do not always want to logically join data at the model level and this provided a nice bridge to filter across explores. It meant we could have one filter dimension listing all the liquid _filters across all explores and if we used this in an explore that didn’t have that filter, then it still worked and that filter would be null. This meant we could have one dimension, rather than lots of variations of that dimension across multiple explores.


We have spent the last 3 weeks trying to get around this through using refinements but looker has now suggested re-architecting our project as refinements will not work either. It seems like the checking is overkill and removes what was a really cool feature that we were sold on in Looker 4. Lookers response to do nothing and cause us to spend months re-architecting a project is completely unreasonable. Personally, looker should re-think this checking to avoid users having to over-complicate lookml or re-design models.


It was also impossible to plan for this as the release notes and changelog completely lacked any detail.

Userlevel 5
Badge

This seems really-really bad. Luckily our instance only broke because we had multiple descriptions to some dimensions 😅

It has been . I get why they are doing this but it’s over-kill and at the expense of a cool feature.

Userlevel 5
Badge

Looks like it has been addressed.


Email from Looker:

" We have discovered and are actively working to address an issue with the introduction of a new liquid parser in Looker 7.10 that surfaces new LookML validator warnings in the IDE. This change has been observed to cause serious performance degradation when validating complex liquid templates.


In the interest of uptime and performance stability, we are updating 7.10 to revert this specific feature only. The rest of 7.10 will remain as published in the changelog, and the new LookML validator warnings will no longer be surfaced.


This redeployment will occur either on Sunday, July 5th or Monday, July 6th depending on your instance’s regional maintenance windows."

Userlevel 5
Badge

One last addition to the release notes - emojies 😅. I think this needs to be addressed as more and more functionality is coming out like this.


💰 - This feature comes with an extra cost ( E.g. Whitelabel menu page)

pngwing.com - This feature only works with Google products ( E.g. Mobility Data Block with BigQuery only)

This is one of things I have been waiting for the longest! The ability to integrate visuals from external libraries EASILY.


Periscope / Mode / DASH all do this natively. Looker has a lot to lose by not catching up with these competitors.

Now looker has advised us they can’t help as apparently looker code has a “warranty”? Anyone ever been told that? With regards release notes, can we add in which sections of code are going to be coming out of “warranty”?

Userlevel 7
Badge +1

Moving this conversation to the other thread where I replied: Lookml Code "Warranty"? Anyone ever been advised that their was a warranty on Lookml?

Reply