Looker 3.50 Release Notes


(Margaret Rosas) #1

Release Rollout Begins: July 5, 2016
Release Final & Download Available: July 14, 2016

Preparing for Release

Please take notice of items marked with a as they indicate changes to existing functionality and may require your attention.

Notable Features

Custom Filters

Custom filters give you full control over filter logic via ANDs, ORs, and parentheses. They leverage the same Looker Expressions used when building Table Calculations.

As of this release, Looker Expressions not only power our Table Calculations, but also the new Custom Filters. In tandem with their expanded role in the product, we made a number of significant improvements to both the usability and functionality Looker Expressions.

Trend Lines (Labs)

Allows the addition of trend lines to cartesian charts using a variety of regressions.

Explore Region Data on the Interactive Map

We’ve made improvements to interactive maps which now support map layers, and we’ve built in map layers for U.S. zipcodes and counties.

Documentation Updates

You can use our new menus (now at the top) to find the page you want. The screenshot below shows the Explore menu.

On each page, the left side lists the contents and enables you to jump to other sections. On the right side, we’ll be gradually adding links to related content.

Legacy Feature Updates

The following 3 features are being turned off and will be removed entirely in 3.50:

  • $$ substitution
  • Joins declared in views
  • LookML “scoping” Parameter

Details on how to migrate off these features can be found here.

Additional Features by Section


Trend Lines [Labs] and Reference Lines

See documentation: Area Charts - Bar Charts - Column Charts - Line Charts - Scatter Charts

  • Added the ability to specify the location of reference line text. Options are Left, Right, Center. Learn more.

  • Options include linear, exponential, logarithmic, and moving average.

  • Moving average type has options for 7, 14 and 28 days.

  • Any number of trendlines can be added.

  • Each trendline can be set to a specific series.

Plot Region Data on Interactive Maps

  • U.S. Zipcode Tabulation Areas and U.S. Counties Map Layers. These new map layers can be used on the interactive map. The zipcode field type automatically works with the interactive map. Learn more.
  • TopoJSON Map Layers on Interactive Maps.
  • All existing TopoJSON map_layers in LookML now work on interactive maps, including built in maps like Countries, U.S. States, and U.K Postcodes, and the new Zipcode and County layers. Learn more.
  • Added a heatmap opacity setting and now map labels can show above or below the heatmap. Learn more.


Custom Filters

See documentation: Custom Filters & Looker Expressions

  • Full control over filter logic via ANDs, ORs, and parentheses
  • Access to same functions available in table calcs
  • Can be used in conjunction with regular filters
  • Contextual help pane

Table Calculations Enhancements

See documentation: Table Calculations & Looker Expressions

Improved Explore URLs

The explore page URL now has a shorter, more stable, format and the URL does not contain filter values or other data. Learn more

Powered By Looker

  • Embedded explore pages now broadcast a Javascript event whenever the page state changes. Enables the ability to build a custom saving mechanism for embed users. Learn more
  • Support for wildcard subdomains in the embed domain whitelist. This makes it much easier to whitelist groups of embedded domains.
  • Fixed an issue where downloading a PDF from an embedded dashboard did not work.
  • Fixed an issue where filter values on embedded dashboards did not wrap properly.


  • Improved unsubscribe flow for scheduled content.
  • Text objects on Dashboards will now appear in Dashboard PDFs.
  • When scheduling content, the To section now accepts comma-separated lists of email addresses.
  • The schedule reassignment modal (accessed via admin/scheduled_plans ) now includes the name of the Look or Dashboard.


  • Added date_start and date_end to date filters This allows the beginning and end dates in a date filter to be explicitly referred to in LookML. This is useful for table wildcard functions in BigQuery or partitioned columns in Hadoop. Learn more..
  • PDT Trigger Check / Regenerator can now rebuild tables concurrently. See documentation.


See dialect support level definitions.

  • Changes to Database Connection Credentials. Looker will no longer pass authentication information through the driver’s connection string and will instead pass credentials as part of the connection properties. Learn more
  • Denodo. Added Level 2 support. See documentation.
  • Oracle. Changes to database connection configuration. See documentation
  • Snowflake. Queries now support timezone conversion. This change will enable features like User Specific Timezones on Snowflake. Learn More.
  • Presto. Added support for PDTs.
  • Vertica. Updated the Vertica 7 connection to Vertica 7.1 (Vertica 7.0 users should use the Vertica 6 connection; to use all Vertica 7 features, upgrade Vertica to 7.1 and connect with that option)
  • Teradata Presto. Added support for version 141t. Contact support for more details on connecting.
  • Hive. Deprecating support for 0.13. This is the last release that will support Hive version 0.13. Starting with 3.52, Looker will support Hive version 1.2+.

Administration & Security

  • Additional Sub-Admin Permissions. Permission to view the PDT panel can be given with the see_pdts permission. See documentation.
  • Usage panel update. Can now be downloaded as a PDF. See documentation.
  • Connections page update. Offer to configure PDT support exists only if the dialect supports PDTs. See documentation.
  • Regenerator based PDT builds now appear in query panel.

General Tweaks and Bug Fixes

  • Restored the ability to color values in Single Value visualizations. See documentation.
  • Model sets on the admin/model_sets/new page will now appear in alphabetical order.
  • Updated 404 Message to let users know that either the page is not found or they do not have access to view it.
  • Fixed an issue that caused filter input boxes to take up unnecessary space.
  • Fixed an issue where scheduled tasks could temporarily be sent from a staging instance.
  • Fixed an issue where text objects caused LookML Dashboards to break.
  • Fixed an issue where scale types didn’t update correctly when switching series positions.
  • Fixed an issue on PostgreSQL with slow relative date filters.
  • Fixed an issue where errors due to ${TABLE} deprecation caused broken dashboards.

The sql_trigger_value check process
(Margaret Rosas) #2


The pivot_where documentation reads:

pivot_where(select_expression, expression)

The pivot_where function returns the value of select_expression for the pivot column which uniquely satisfies select_expression or null if such a column does not exist or is not unique.

I haven’t had the opportunity to try this function out yet, but I assume it should read something like the following, no?

The pivot_where function returns the value of expression for the pivot column that uniquely satisfies select_expression or null if such a column does not exist or is not unique.

(Tig Newman) #4

Hi Michael,
Yes, you are correct. As you say, it should be simply expression so it becomes:

The pivot_where function returns the value of expression for the pivot column which uniquely satisfies select_expression or null if such a column does not exist or is not unique.

I’ll get that fixed in the documentation and thank you!

(Richard) #5

Very excited about moving average trend lines and custom filters, fantastic update!

(john) #6

any update to pdf download performance?

(Rjnordeen) #7

Or any update on being able to download image graphic directly?

(Carter Moar) #8

We’ve got an active project that should be wrapping up in the next release or two (Looker 3.52 or 3.54) that will make Looker smarter about how it runs the PDF process (it won’t try to render a PDF until the query returns results, it’ll be smarter about handling situations where users mash the download button, etc.). I believe that the project will also introduce a rendering event queue, so that you won’t end up in a state where you’re trying to render a dozen dashboards simultaneously.

I don’t believe that this is slated to come in quite as soon, but it’s certainly on the radar.

(Carter Moar) #9

(Parijat Talkad) #10

in regards to the Heat Map, will there be the ability to have a graphical representation of data where the individual values contained in a matrix are represented as colors? (Not just a geo location map but a matrix)

(Dave Hunsinger) #11

Are you referring to a square heatmap or a treemap that represents varying x/y values versus a georeferenced heatmap?

(Parijat Talkad) #12

the square heatmap!

(Dave Hunsinger) #13

Ok - just making sure! That’s on the Product team radar, but I do not know what release it will land in at this time.

(Abby West) #14

This will be in 3.54, which we’re cutting this week!

@rjnordeen - we’re expecting to offer this feature with the release of 3.56 or 3.58.

(john) #15

Thank you @abbywest. didn’t see pdf download performance improvement in the 3.54 release notes, could you confirm that it made the cut? can’t wait!

(Abby West) #16

Hi @john_handshake - it is there, under a slightly different name …

Resource Manager for Scheduled and Downloaded Tasks
By default, the system will process downloading and rendering tasks in pairs. Instance administrators will see a new source in the query panel called “renderer”.
These changes should result in fewer timeouts on rendered content and reduce impact on the Looker instance.

As part of this change, we made a bunch of backend improvements to how we render pdfs for downloading and scheduling!