Looker 5.0 Release Notes

release

(Carter Moar) #1

Anticipated Deployment Dates

Release Rollout Begins: October 8, 2017
Release Final & Download Available: October 19, 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

LookML Collaboration

Your LookML developers can easily manage and organize multiple simultaneous pieces of work, making development easier than ever. In addition to their dedicated development mode, users can swap to or create new git branches dedicated to specific tasks. Working on a major project but need to fix a typo real quick? No more do you need to abandon or push your intermediate work.

Lookless Dashboards

Dashboards no longer have to be tied to existing content, but instead can be made of Tiles which can be built directly on the Dashboard. If a Look exists only to populate a Dashboard you can replace it with a Tile that lives on the Dashboard rather than in Spaces, simplifying your content system.

Landing Page

A new way to find what matters most to you! Looker 5.0 introduces a new landing page with highlighted favorited and suggested Looks and Dashboards, providing a launching point for your new users and surfacing content you might have overlooked. This is currently an experimental Labs feature, so you can bank on fun new additions in the future. Learn more.


Features by Section

LookML and Development

- **LookML Collaboration.** Added the ability to create development branches that are not specifically tied to any individual's development mode. Use these branches to work with others on a project or to separate discrete pieces of work you have in progress. [Learn more.](https://docs.looker.com/data-modeling/getting-started/version-control-and-deploying-changes#working_with_git_branches) - **Reference pivot values in Liquid.** You can now reference the `{{ value }}` of a pivot value without it being the one you're actively clicking on. - **Timezone support for Native Derived Tables.** Use the `timezone` `explore_source` parameter to convert to a consistent timezone or, for ephemeral derived tables, to the timezone your query is in. - **Native Derived Tables are no longer a Labs feature.**

Content Management and Discoverability

  • Lookless dashboards are out of beta. Add Tiles to your Dashboards either by creating the query directly on a Dashboard or saving from an exploration.
  • :alembic: Landing Pages. [Labs] Introduced a new, experimental home page that will surface Favorite and Suggested content as soon as you log in.

Scheduling and Downloading

  • Increased allowable schedule size. Attachment-only formats (currently txt, csv, html, md, and json) support scheduling without any row limit applied. Keep in mind, though, that emails of more than 7.5mb may be returned to the sender. Learn more.

Dialects

  • Druid. Introduced initial support for Druid. Learn more.
  • :zap: Google BigQuery Legacy SQL. Updated the name from “Google BigQuery” to “Google BigQuery Legacy SQL” in the connection panel, SQL Runner, and other locations the dialect name is displayed.

Platform and Administration

- :zap: **Disabled and removed a number of Legacy Features.** [Learn more.](https://discourse.looker.com/t/legacy-features-end-of-life-schedule/2067)

Security

- Better protect against certain authentication replay attacks. - Addressed a file disclosure vulnerability. - Addressed a file system vulnerability. - Identified and addressed a logging issue.

Dashboards and Visualizations

  • :alembic: Data Merge [Labs]. Introduced the experimental Merged Results Labs feature, which allows users to visualize results from multiple data sets without modeling. The ability to save as well as add to Dashboards and Looks is currently in development. Learn more about using Merged Results. Learn more about opting in to the experimental beta group.
  • :zap: Pie chart’s labels will no longer display for every value if there is not enough room. However, labels will no longer get cut off.

General Tweaks and Bug Fixes

  • Dashboards, Visualizations, and Explore
  • Fixed an issue where filter values would not apply on a Dashboard unless the user hits enter or exits the input box before running the query.
  • Addressed an issue where users could Clear Cache & Refresh even when Required Filters were not entered.
  • Fixed an issue where explores were not displayed using their label.
  • Fixed an issue where a blank page was displayed when a user did not have access to a Look.
  • Fixed an issue where errors would occur when results have arrived but Table Calculations haven’t been performed.
  • Fixed an issue where adding a Look to a Dashboard resulted in an ineditable, incorrect title on your Dashboard.
  • LookML and development
  • Fixed an issue where Looker incorrectly errored on parameters that carried through includes.
  • Fixed an issue where a view file with a parameter declaration could not be extended.
  • Fixed an issue related to parameter and Liquid joins.
  • Looker now properly errors when percentiles are used on dialects that don’t support them.
  • Fixed an issue where the IDE could display false error messages.
  • Fixed an issue where the development mode toggle did not display on Admin pages.
  • Improved the error message when Liquid references were not properly scoped in Derived Tables.
  • Fixed a few issues with the Don’t Revalidate Unchanged LookML Labs feature.
  • Content Management and Discoverability
  • Fixed an issue where Dashboard and Look lists did not load at the same time, making it appear as though the page had not finished loading.
  • Fixed an issue where the Model column was not displayed in the LookML Dashboards space.
  • Fixed an issue where you couldn’t use the Content Validator to replace spaces ().
  • Fixed an issue where LookML Dashboards with numeric titles broke search.
  • Dialects
  • Google BigQuery Standard SQL. Fixed an issue that produced a SQL error when Symmetric Aggregates where applied to filtered measures.
  • Google BigQuery. Fixed an issue where canceled queries displayed a status of “error” in the Query panel. This applies to both Google BigQuery Standard SQL and Google BigQuery Legacy SQL.
  • Addressed some issues with query killing on Presto, Spark, DB2, Qubole, Redshift, and Teradata.
  • Platform and Administration
  • Fixed an issue where the API did not properly report errors for duplicates in query.slug to callers.
  • Fixed an issue where new user emails did not send out.
  • Fixed an issue where the Query panel could still display orphaned queries.

Modeling Conversion Rate
(Josh Siegel) #2

Does this actually break anything or is it just a display name thing?


(Carter Moar) #3

It’ll just be a display thing.


(Carter Moar) #4

For information on setting up the Data Merge Labs feature, check out the Learn More video, about 4:15 in!


(Ryan Johnson) #5

What about Data Actions? Are they not in this release? Can you tell us more about the Asana Actions?


(Carter Moar) #6

You can leverage existing Data Action functionality to build your own Asana action and, over the course of the Looker 5 life (though not specifically in Looker 5.0) we’ll be building out more actions for the Action Hub. Right now the only action in the Hub is the Segment one but stay tuned, we’re hoping to have more information available on that front in early December.


(Tig Newman) #7

As @cutler mentions, we have a video on merging results. We also have Merge Results officially doc’d now:


(Camille Vernon | Data Operations Analyst) #8

Is there a way to get the Content Validator to validate against the state of the projects in ones development mode for specific branches? Or does the Content Validator only look at changes made in the branches one owns starting with “dev-”.

I noticed when I made a new branch from master and ran the Content Validator, it was validating changes in my main dev branch. Perhaps I need to branch from my dev branch instead?

Thanks!


(Carter Moar) #9

The Content Validator should be running against your active branch, regardless of what mode you’re in. So if you’re out of dev mode it’ll run against your instance’s production environment and if you’re in developer mode it ought to be running against whatever branch you have active.

If that’s not happening (which sounds to be the case), could it be that you had a few tabs or windows up with different branches? I’ve definitely lost track of which was the “currently active” one myself. If that’s not it do you mind reaching out to support@looker.com or chat so that folks can dig in a little deeper?


(Camille Vernon | Data Operations Analyst) #10

Indeed, I only had tabs open to Develop (in the feature branch, branched from prod/master, which was not my personal dev branch) and Explore with my changes, and the Content Validator was showing me that my personal dev branch was behind Prod (hadn’t done a pull in a while there). I’ll email support@looker.com so they can look into it.

Thanks!


(Timothy Burke) #11

Could you provide an example or a little more description about this feature?


(Timothy Burke) #12

Cross-posting here what I noted elsewhere:

The new landing page is likely great for some, but when the content you’d like to curate is documentation or LookML dashboards, this isn’t ideal. I’d love to see flexible options for home page configuration: specific markdown files (full control for admins) OR Favorite / Popular content (crowdsourced / user-controlled)


(Brendan Vaughn) #13

Does Looker 5.0 still require the installation of PhantomJS for on premise installations?


(philip.martinelli) #14

Hey there @TimothyBurke,

I drafted up an example about referencing pivot values using liquid.

Say I have the following explore and I want to create a drill path to another explore query filtered by the pivoted dimension of the measure I click:

I can define the measure with a link parameter like so:

measure: all_users {
type: number
sql: ${count_users_with_orders} + ${count_users_without_orders} ;;
link: {
label: “pivot test-explore”
url: “https://self-signed.looker.com:9999/explore/ecommerce/users?fields=users.state,users.count_users_with_orders&f[users.gender]={{ users.gender._filterable_value}}”
}
}

When I click a measure value under one of the pivoted dimensions, the dimension’s value is passed into the URL as a filterable value. Say I click on a value under “f”, I drill to this:

I hope that helps!


(Carter Moar) #15

If you’re using PDF or image exports (downloads, schedules, etc.) then, yeah, you’ll need to have PhantomJS 2.1.1 installed on the machine hosting Looker.


(Calvin Lacock) #16

Any idea when the Viz Blocks will be available?


(Carter Moar) #17

The current tentative target is for them to be available for Looker-hosted instances in about a quarter or two.


#18

Looking forward to the release. Do you have a sense of when SQL Runner visualizations will be released? Thanks!


(Carter Moar) #19

The SQL Runner visualization project is targeted for some time during the Looker 5 series, but the folks working on that are devoting more resources toward polishing up Merged Results and getting that out of beta.


#20

great thanks