Looker 4.18 Release Notes

release

(Carter Moar) #1

Anticipated Deployment Dates

Release Rollout Begins: July 16, 2017
Release Final & Download Available: July 27, 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.

Preparing for Release

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


Notable Features

Native Derived Tables

Creating a Derived Table just got a whole lot easier. Rather than being defined with simple SQL blocks, they can now reference existing fields and are defined in LookML. Writing and managing Derived Tables with complex metrics is a snap since you can reference the logic you already have in existing dimensions and measures. Tables will easier to read and reason about and can be SQL dialect-agnostic (although you can still define custom SQL).

More Curated Drill Experiences

Guide your users on a journey! With a new Liquid parameter, you can pass the filters applied to a query through to linked queries, Looks, and Dashboards. Using _filters['view.field'] you can ensure that a Dashboard’s date range applies to deeper explorations, that categorical filters persist, or have users refine what’s presented–say with linked Dashboards that flow in granularity from country to region to city. Learn more.


Features by Section

LookML

  • Native Derived Tables. :alembic: [Labs] Derived tables can now be composed with LookML rather than being defined with a simple SQL block. Learn more.
  • Introduced Liquid parameters to identify a field’s presence in a query. Use _is_selected, _is_filtered, and _in_query to build dynamic behavior based on the fields and filters of a query. Learn more.
  • :zap: _in_query now returns true if a field is in the data table, the filter set, or is being automatically included with required_fields or similar parameters.
  • Introduced the _filters['view.field'] Liquid parameter. This allows you to grab a query’s current filters, especially useful for cross-Dashboard linking as well as hyperlinking to Looks, queries, or other link values. Learn more.
  • :zap: Dashboard, Look, and explore links can now open in the same page or iframe. link values with absolute URLs will continue to open in a new tab while root-relative ones will remain in the current iframe or browser tab.
  • Added the ability to declare a description on LookML Dashboards. Learn more.
  • The LookML generator now automatically applies map_layer_name to country fields. Learn more.

Looker Platform and Administration

  • Integrations
  • Introduced an integration with Segment that will allow you to schedule data to be connected to third party applications for automating workflows. Learn more.
  • SQL Runner
  • :alembic: [Labs] Introduced the ability to search for all of the fields, schemas, and tables on a given connection. Learn more.
  • Filters
  • Suggestions to now match partial words and rank suggested filter values based on relevance.
  • API
  • Project management. Added endpoints to configure a LookML Project’s Git settings and generate and fetch public-private SSH key pairs.
  • Dashboards. LookML dashboard_elements can now be directly associated with a query rather than be routed through a Look
  • Dashboards. dashboard_elements now have a title_text parameter that will override the Look title or will provide a title if there is no associated Look
  • Instance Administration
  • Legacy Features. Updated the Legacy Feature program to provide greater clarity into deprecation timelines and to add deprecation releases for existing Legacy Features. Learn more..
  • Authentication. Updated the SAML setup page to clarify the IdP Audience field’s relationship to the SP Entity ID.
  • :zap: **Permissions.**The Rebuild Derived Tables & Run query option will now only be available to those with the Develop permission. Learn more..
  • Schedule Management. Updated the Scheduler History Admin panel to include information whether or not the job was a test and the type of scheduled job it was. Learn more.

Dialects

  • Amazon Athena. Query killing is now supported.
  • Netezza. Added initial support for Netezza.

General Tweaks and Bug Fixes

  • Scheduling and Downloading.
  • :zap: The “Edit Schedule” gear menu option now says “Add Schedule” if no schedule exists.
  • :zap: Added an identifier to files sent to S3 and SFTP to ensure file names are unique.
  • Users can now download data from the Drill modal. Learn more.
  • The Scheduled Plan Dialog now closes with the ESC key
  • Fixed an issue where email addresses removed from a schedule were still listed on the External Emails Admin page.
  • Fixed issues related to zooming on PDF versions of maps.
  • Fixed an issue where scheduling to SFTP could send an empty file.
  • Fixed an issue where Automagic Heatmap Map visualizations could generate empty PNG files.
  • Fixed an issue where empty Totals rows could cause an export to fail.
  • Fixed an issue where “View this data in Looker” links in scheduled emails could lead to an invalid page.
  • Fixed an issue where visualization issues could result in an empty CSV download.
  • Dashboards and Visualizations.
  • The vertical order of series in legends now matches their order in stacked Column visualizations.
  • Fixed an issue where stacked Column visualization value labels did not properly hide.
  • Fixed an issue where tooltips did not appear on Area visualizations.
  • Fixed an issue where Timeline visualizations rendered blank due to invalid value format values.
  • Addressed a number of issues related to having multiple maps on a Dashboard.
  • Fixed an issue where clicking “View Access Settings” from a Dashboard’s gear menu lead to an endlessly spinning page.
  • Spaces and Content Management.
  • When LookML Dashboards are imported into a Space, the banner notifications now link to the Dashboard and Spaces to which it was moved.
  • Fixed an issue where a user was displayed the Explore From Here link in Spaces despite not having the permissions to do so.
  • Fixed an issue where, after moving a Dashboard, its Looks remained in the departure Space until the page was refreshed.
  • Instance Administration.
  • Fixed a bug where Groups could not be added to an individual user’s Space in a closed system.
  • Fixed an issue where SAML users’ authentication errors were not appearing on the login page.
  • Fixed an issue where SAML, LDAP, or Google Auth users could be sent password reset emails despite having no password to reset.
  • Google BigQuery. Fixed an issue where Max Billing Gigabytes entries could fail to validate.
  • Fixed an issue where grouped Explores did not sort correctly.
  • Fixed an issue where the browse Spaces icon was missing in embedded contexts.
  • Fixed an issue where users could not edit their email preferences.
  • Fixed an issue where Access Filters in Liquid conditions caused erroneous errors.
  • Fixed an issue where certain timeframes produced erroneous to allow_fill warnings.
  • Fixed an issue where the Help menu’s search bar was unresponsive.
  • Fixed an issue where markdown tables could render with very narrow columns.
  • Fixed an issue where the pivot_where Table Calculation would erroneously error.

Filter suggestions are not sort based on order_by_field in 4.18
How to manually rebuild persistent derived tables
(Carter Moar) #2

#3

very excited about testing the Native Derived Tables functionality!


(Carter Moar) #4

me too! I haven’t had much opportunity to do much with them yet, but I’m really looking forward to simplifying what I’ve got


(Tig Newman) #5

@mplooker – Lloyd has just posted an article about using native derived tables here –


(Scott Schaen) #6

Some nice updates, but I have one concern…

We have a lot of non-dev users that rely on “Rebuild Derived Tables and Run”…Is this going to be optional? This would increase reliance on our few Looker devs, which defeats the point of Looker. We’re using BigQuery where rebuilds are expensive; this is why derived tables are run on-demand as opposed to scheduled.

I might have to hold off on updating if this is not available as an option :-/


(Colin) #7

Hi Scott,

We don’t anticipate any changes here, but swapping your PDTs to persist_for (or using DTs without persistence) should handle any needs for derived tables to be rebuilt on-demand. Some combination of persistence and triggering could also be baked into a datagroup if you want these built regularly as well.

Colin


#8

thanks, @tig; this is very helpful!


(Zian Sun) #9

any details regarding Liquid parameters to identify a field’s presence in a query?


(Scott Schaen) #10

Hi Colin,

Neither options satisfy our use-case :-/

To confirm, only “developers” will be able to select the “Rebuild Derived Tables and Run” option? Is this not definable in permission sets?

Thanks,
Scott


(Colin) #11

Hi Scott,

Sorry to hear that won’t work well. Access to PDT regeneration has broad implications on instance performance, so we wanted to better align the ability with an understanding of the development. There may be some future state where we create more intermediate permissions, but we’ve strove to keep the number of permissions to understand small and manageable, so we try our best to avoid permission bloat.

If you want to shoot me a note to colin@, we can talk more about it directly to understand your case better.

Colin


(Michael King) #12

Has this release changed the behaviour of filters? Previously we were able to double-click a filter after application to return it to an editable state, but this seems not to be possible as of this release.

We were also able to select and copy applied filters (for the purposes of pasting it elsewhere, for example), and this also seems to not be working.


(Frank Kutzey) #13

Same here, copy & paste doesnt work anymore + I can see all filter values if multiple are selected.
Is that a known bug already ?


(Michael King) #14

I’m also used to hitting Ctrl+Enter to run an explore, but this no longer works as expected when your cursor is in the text field for a filter.


(Tig Newman) #15

Hi Zian Sun,

You can learn more about this on a new page – Liquid Variables.

That page does not yet show in the menus yet (or in this article’s “Learn More” for that item) because it has links to not-yet-rolled-out sections in other pages. So… if you forgive us that not all of the links are ready, you can take a sneak peek. There are 3 sections on that page that have info or examples for the three variables related to a fields’s presence in a query:

We’ve rolled out the related changes so the links on this page are ready too!


#16

We are using a new UI library for filters in this release and it hasn’t yet reached feature parity with the previous implementation. We are working on supporting copy/paste and should have a fix available soon. Until then, as a workaround, try switching the filter type to matches advanced to copy/paste. Hope this helps.