Looker 4.8 Release Notes

  • 16 February 2017
  • 25 replies

Userlevel 5

Release Rollout Begins: February 20, 2017

Release Final & Download Available: March 2, 2017

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.

  • Scheduling and Downloading

  • Dashboards and Visualizations

  • Looker Platform and Administration

  • Content Management and Discoverability

  • LookML and Development Environment

  • Dialects

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.

  • Changed the location of deploy keys for newly created projects.

  • “Rebuild Derived Tables & Run” option on Explore pages is now hidden for all embedded users.

  • Moved Public Access settings for Looks to the Look Settings modal.

  • Disabled the Save button from the Edit Look modal until the query is finished updating.

Notable Features

Conditional Formatting

Make the results that matter pop! Table visualizations now sport a Conditional Formatting labs feature; use it to define rules that color code values that are low, high, or otherwise of interest.

Filterable Dashboard Schedules

Dashboards schedules can now each have their own filter values, allowing you to customize results for your recipients. Send regional teams results filtered for their territories of interest and filter weekly, quarterly, and annual reports independently. Use a single Dashboard to manage your scheduling needs!

Features by Section

Scheduling and Downloading

  • Filterable Dashboard Schedules. Introduced the ability to change filter values for each schedule associated with a given Dashboard.

  • Inline Tables is now the default email delivery option for new scheduled content. Existing schedules won’t be changed.

  • Added API support for downloading Dashboard PDFs.

Dashboards and Visualizations

  • Conditional Formatting [labs] for Table visualizations. Learn more about Labs features.

  • Added drop-down options to the Add/Edit Dashboard Filters interface.

  • Trend Line equations and R2. Trend Line labels can now display the equation used for the line or the Trend Line’s R2 value.

Looker Platform and Administration

  • Powered by Looker

  • Saving from an Embedded Context. Introducing the ability for embedded users to save content the ability to curate content for embedded users.

  • Scheduling from an Embed Context. Embedded users now have the ability to schedule Looks and Dashboards.

  • Embedded Look URLs can now be modified to override filter values.

  • JavaScript embed events can continue firing across page loads.

  • Exposed embed-related external_id fields in already-existing credentials_embed and space routes.

  • ⚡ “Rebuild Derived Tables & Run” option on Explore pages is now hidden for all embedded users.

  • Changed the location of deploy keys for newly created projects. Looker-hosted instances will not need to take any action. Action is required for customer-hosted clusters creating new projects. Learn more about syncing git SSH keys. Learn more about shared storage clusters.

  • Connection tests. Added an option in the Connections Panel to test all connections or each individually.

Content Management and Discoverability

  • Unused Content and Trash. Introduced two sections that allow instance Admins to, in bulk, clean up all of the unused and unneeded content on an instance.

  • The Unused Content page will pull together all content that hasn’t been touched in 90 days. Admins can then delete those items, which will move them to the Trash.

  • Trash. Items in Trash can be restored or permanently deleted at the Admin’s discretion.

  • Tables in Spaces are now sortable and are paginated. This does not apply to the LookML Dashboard space.

  • Move and Copy Modals in Spaces now match the new Save Look modal.

  • Run on Load for Looks. Introduced the ability to set whether Looks run immediately. This setting can be configured in Edit Look Settings.

  • Moved Public Access settings for Looks to the Look Settings modal.

  • It is now possible to delete a Dashboard from the Dashboard page.

  • Markdown file navigation. Added the ability to create a navigation menu for Markdown documents.

LookML and Development Environment

  • Fiscal years and quarters. Introduced a fiscal_month_offset LookML parameter to allow for custom fiscal years. Learn more

  • Better SQL Runner error highlighting.

  • Projects can now include JSON files.

  • Updated the syntax for map_layers so that they can now reference JSON files included in the project.

  • Introduced the ability to add files to a project by dragging and dropping them into the LookML IDE.


  • BigQuery Standard SQL. Improved the LookML project generator.

  • BigQuery Legacy SQL. Removed EACH from GROUP BY, improving performance for certain queries.

  • Presto. Added support for Qubole’s Presto as a Service for Presto versions 0.157 and 0.142.

General Tweaks and Bug Fixes

  • ⚡ Disabled the Save button from the Edit Look modal until the query is finished updating. This addresses an issue where Look edits would not persist if saved too quickly.

  • Fixed an issue where menus items that ran off the page could not be reached by scrolling.

  • Removed the Go to LookML link for users who are not able to see LookML.

  • SQL errors are now only exposed to users with the see_lookml permission.

  • Fixed an issue where the SQL tab was blank on Safari.

    Fixed an issue related to escaping the \ character in filter values.

  • Fixed an issue where Reset Look did not reset query results.

  • Fixed an issue where webhook delivery could block scheduled content delivery.

  • Fixed an issue where Single Column PDF Dashboard element titles were mixed up.

  • Fixed a issue where the ‘Create Schedule’ option was not available in the Details pane of the Look page.

  • Fixed an issue related to downloading results from drill overlays.

  • Fixed an issues that resulted in Select 10 SQL Runner functions not working properly.

  • Fixed a SQL Runner issue related to listing tables in on Spark 2.0.

  • Group labels for measures are now the same color as measure names in the field picker.

25 replies

Userlevel 3

Awesome release, thanks guys.

Userlevel 3

People are gonna love conditional formatting!

pretty sweet stuff!

please consider HTML rendering to work when “Download…” results!

thank you!

Userlevel 2

Several very helpful updates. Looking forward to getting the release!

Userlevel 1

What do these entail?

Userlevel 2

Great update! One question - will filterable schedules be available at some points for looks - why only for dashboards? Thanks!

Userlevel 5

  • Rather than strictly relying on the advanced filter expression syntax, the Edit Dashboard Filters modal now uses the drop-down interface, just like explore filters.

  • Events will continue to fire regardless of what you do in an iframe, where previously navigating to a new page stopped those events

  • You can now have .json files in your projects, just as they’d previously had .lookml, .lkml, and .md files. This allows map_layers to reference those topoJSON files.

  • You may have noticed gear menus and such partially populating off-page, making some menu items unreachable–that shouldn’t happen any more (probably my favorite bug fix of the last three releases).

  • There was an issue on some SQL dialects (certainly MSSQL, no other dialects to the best of my knowledge) where \s entered as filter values were not being escaped.

can kinda do this now with access filters by recipient

Userlevel 2

True, but that’s predicated on the intended recipient being a Looker user which isn’t always the case. That also doesn’t take into account if we wanted to send different filtered versions of the same underlying look to the same user.

Userlevel 1

Hey Dan - we are adding filterable schedules to Looks. Targeting next release (4.10) to land that feature. We rolled it out for dashboards first, but are definitely intending for it to work with Looks as well.

Loving the new stuff in this release.

Question about Conditional Formatting… I realize this is a labs feature and probably has more coming, but is there a plan to be able to apply different conditional formatting rules (and/or different calculation of the scale) for different columns? I was attempting to use conditional formatting in a table that has both a user count column which can go up into the hundreds of thousands AND a separate column that has percentages. It seems to be applying the same scale/color rules to both so that percentage column is always all red since I think its being interpreted as either all less than 1 (if its using the unformatted value behind the scenes) or all less than 100 and being colored against the max/min values of the user count column.

Does this question make any sense? 🙂

Userlevel 1

Hey Max - yes, this definitely makes sense. You are right, we made conditional formatting a labs feature so that we can collect feedback and consider adding to the feature. We wanted to make it available as quickly as possible, so we released without the full potential feature set. Being able to apply scales on a column by column basis is one of the enhancements we are considering, in addition to applying different rules to different series, which would accomplish the same thing in your use case.

To confirm that you have seen this, you can choose which series the rule applies to by expanding the series you want to disable and toggling “Apply Conditional Formatting.” I understand this doesn’t solve your full use case, but it should mitigate the problem with the percentage column impacting the scale and being colored all red.

Thank you for sharing your use case and feedback. It’s very helpful for us in determining next steps for this feature.

Just downloaded 4.8 and wanted to validate a couple of assumptions on the conditional formatting feature.

  • Does it only apply to measures? Some of the formats (null/not null, greater than value, less than value) would seem to be applicable for dimension columns as well. An example here would be i__looker/look explore where there is a “deleted time” column, it would be good to highlight that value when it’s not null. (I have an “All Looks In Looker” Look for keeping track of what gets created in our instance)

  • For the future, can the conditional format be based on a value but applied to a different column. One potential use that I could see for this feature is to use it with a table and highlight either the main text value (like Category Name in the example) when the value of a specific measure is above or below a threshold value.

  • Right now it’s limited to rules applying for all measures, but the series editor allows you to define multiple rules, what’s the intention there, and what happens if multiple rules apply to the same value? My assumption is that the rules would take effect in the order they’re defined, but that still leaves the question of which one sticks (first or last)?

Userlevel 1

Hey @Robert_Ellman1

It currently applies to both dimensions and measures, but only numeric columns right now. If more than one rules apply to a value, then only the first rule will take effect and the others are ignored (the first one sticks).

Thanks for your feedback on this feature! This is just an initial rollout now, but we’re hoping to add more functionality down the line. I’ve brought these suggestions to our product team.

BigQuery says “Use GROUP EACH BY when your query has failed with a resources exceeded error message.” – is there a way to do this if we get an error in Looker?

Hey @joshsiegel, Thanks for the question. We found that EACH with the GROUP BY for Big Query Legacy was causing more problems than solutions. In some cases it was causing queries to run twice and also slowed performance. Because of these and the other issues with EACH we have decided to not support it.

Hi, is there a way to remove or disable “Save as a Look…” option from the menu of embedded (with sso api) explore screens?

@Alexander_Kartavenko What you’re looking for is the save_content permission. Deny that permission to any user and they will not be able to save Looks or Dashboards

@William_Lane I’ve already tried this, and even when I do not pass save_content permission “Save as a Look…” option is still in the menu, am I missing something?

@Alexander_Kartavenko That sounds strange! Send an email to help.looker.com and we can look into this with you

Very excited about this release! Two questions regarding features:

For conditional formatting, I noticed you are planning on enhancing conditional formatting to apply on a column by column basis - will this include pivoted fields? It would be really great to be able to format a measure based on the pivot, rather than 1 formatting rule applying to all cells.

For scheduled dashboard filters: I’m already using this feature for teams, which is awesome, but I noticed that the link ‘View this data in looker’ does not apply the same filter as is in the email, it take users to the default dashboard. Is ther any way to make that link also filtered so teams don’t get confused when the inline email has a different filter than the dashboard link?

Thanks so much guys!

Userlevel 2

Hi @kdietz,

Glad to hear you are enjoying the new features in Looker! On your question about including pivots, currently in the first iteration of a column by column basis, pivots will not be supported. The engineering team is currently working towards a way to best go about including pivots. For the schedule, are you referring to the values in the filter? That when you click on the link, instead of the values in the schedule, they are set to the default values of the dashboard?



Hi @vincent , thanks for the quick response! For pivots, that’s great to hear, happy to have the labs feature out, and looking forward to future improvements.

For dashboard, yes that’s right. The email goes out to users with the correct dashboard filter applied, but when they click the email link to go to the dashboard in looker, it takes you to the default filter, rather than the scheduled filter (which confuses some users).

Userlevel 2

Hi @kdietz,

That is expected behavior as of right now since all we do is include the link to the dashboard without supplying any of the filter parameters, just the dashboard ID. I can, however, loop in my Product team about this.



Thanks @vincent! Good to know the behavior is expected, and I definitely think it would be useful to loop in the Product team. Adding the schedule filter to the inline URL would be a huge plus to this feature.