Looker Connect Training

Help Center

Documentation

Community

Cloud Certifications

Technical Tips & Tricks

Previously internal product knowledge now available for you to explore!

  • 1,744 Topics
  • 406 Replies

1744 Topics

molly.lippsett
Looker Staff
molly.lippsettLooker Staff
published in Technical Tips & Tricks

HADOOP—Error: Error reading from hdfsKnowledge Drop

Last tested: Aug 13, 2018What does this mean when you see it in a database connection test? It's possible for blocks of the Hadoop File System to get corrupted - in this case you'll see a message like this. Removing the corrupted block from hdfs removes the error. This content is subject to limited support.                  

430
molly.lippsett
Looker Staff
1 year ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

Can I reuse my LookML code? (Reusing Code with Extends)Knowledge Drop

Last tested: Jan 21, 2019 For more information on Extends and how to reuse your LookML code, review the documentation here.But how does Extends work? What if there is the same parameter on my new file?At a high level, Looker performs a object-by-object comparison and the new file takes precedence if the same object exists on both files. The steps are:COPY: A copy is made of the defined LookML object that is being extendedMERGE: The copy is merged with the defined LookML object that is doing the extendingRESOLVE CONFLICTS: If a LookML object is defined in both places, the object doing the extending is usedFINISH: The new LookML object can be used within the model just like any other LookML object This content is subject to limited support.                  

1040
Department of Customer Love
Looker Staff
1 year ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

Pairing IdP Attributes with Looker User AttributesKnowledge Drop

Last tested: Oct 3, 2019 You can pair IdP attributes to Looker user attributes in order to get user-specific settings into Looker (for example if you're using a parametrized connection or access grants). This can be helpful if you already have attributes set in IdP that you want to use in Looker, or would prefer to set permissions like this in IdP rather than Looker, so the IdP can act as the single source of role truth.ExampleFor example, you might have a SAML attribute set in your IdP for department which takes on the values finance, sales, or marketing. In Looker, you may require only the finance department to see the payroll values of the organization.In order to make oversight of this user attribute more seamless, you can map a Looker user attribute to the department SAML attribute. Here are the steps:Create a new user attribute in Looker. (If you have an existing user attribute in Looker that you want to map to the SAML attribute, then you can skip this step!) In this example, we

1140
Department of Customer Love
Looker Staff
1 year ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

Why is my NDT showing a "No such field ' ' in source explore ' ' " error?Knowledge Drop

Last tested: Mar 14, 2018 This has been shown to happen to NDTs for 2 main reasons. One is a Typo in the view_name.field_name reference and the other reason is a missing fully scoped field reference.Typo in field nameError Message:A LookML model issue occurred.No such field "produtypo.category" in source explore "order_items"bucket1/ndttest.view.lkml:4 NDT LookML:view: ndttest {derived_table: {explore_source: order_items {column: category { field: produtypo.category }column: count { field: products.count }Where there is a typo at productypo and that should be productsMissing Full Scoped Field referenceError Message:A LookML model issue occurred.No such field "category" in source explore "order_items"bucket1/ndttest.view.lkml:4 NDT LookML:view: ndttest {derived_table: {explore_source: order_items {column: category { field: category }column: count { field: products.count }  Where field: category should be written fully scoped as field: products.category  This content is subject to limite

3740
Department of Customer Love
Looker Staff
1 year ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

Generate Date Series/ Create a calendar view of future datesKnowledge Drop

Last tested: Aug 26, 2019Some dialects have a generate_series function that makes this trivial. We have a Community article that explains a workaround you can write in SQL for MySQL, Redshift, BigQuery and Snowflake.For example, Redshift does not implement generate_series(), so we will use the row_number window function over a bunch of cross joins to generate a numerical sequence, and then add the redshift date logic:SELECT '2015-01-01'::DATE + d AS dateFROM ( SELECTROW_NUMBER() OVER () -1 as dFROM(SELECT 0 as n UNION SELECT 1) p0,(SELECT 0 as n UNION SELECT 1) p1,(SELECT 0 as n UNION SELECT 1) p2,(SELECT 0 as n UNION SELECT 1) p3,(SELECT 0 as n UNION SELECT 1) p4,(SELECT 0 as n UNION SELECT 1) p5,(SELECT 0 as n UNION SELECT 1) p6,(SELECT 0 as n UNION SELECT 1) p7,(SELECT 0 as n UNION SELECT 1) p8,(SELECT 0 as n UNION SELECT 1) p9,(SELECT 0 as n UNION SELECT 1) p10) This will give you row for every date from Jan 1, 2015 to Aug 9, 2020. Adding an additional (SELECT 0 as ...) will double

16430
Department of Customer Love
Looker Staff
1 year ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

Why is an image (PNG / PDF) download request stuck on "Queued for Rendering" for a long time (20+mins)?Knowledge Drop

Last tested: Jun 4, 2019 Since we can only run two simultaneous rendering processes by default (see the --concurrent-render-jobs=<i> startup flag), two slow queries can back everything else up. So, image download jobs really can be queued for rendering.Looker will try to run the queries and cache the results before entering the renderer queue, but sometimes the cache will be invalidated by the time Looker starts rendering. Below is an explanation and image of the behaviour seen.~30mins to fully render on the EIn that example, the instance had an Explore where a newly run query could take explore page. When this instance had multiple queries run based on that underlying Explore, it caused a render backlog. Even the simplest PNG download request had to wait for that ~30min explore to finish rendering (or have it be manually stopped via Admin > Queries Panel)if you see a long-running Queued for Rendering page, review the Queries Admin Panel for long-running queries. This can help

2120
Department of Customer Love
Looker Staff
1 year ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

Advanced LookML - Cohort and Retention AnalysisKnowledge Drop

Last tested: Jun 24, 2019 What is a Cohort Analysis?Read this data tutorial here (5 minutes) and check out this introduction video to learn about the fundamentals about cohort analysis: What’s the point of Cohort Analysis?The goal of a cohort analysis is to analyse the performance of cohorts relative to each other and over time. Once we understand how different cohorts are performing, we can then take action.What exactly is a “cohort”?In a nutshell, a cohort is simply a subset of users grouped by common characteristics. In the context of BI and SaaS, a cohort usually refers to a subset of users specifically segmented by some key date (i.e. the first time they visit your website, the date they perform a specific action, the date they made their first purchase/registered, etc.). Based on this key date, we can then group users together by the week / month, or whatever timeframe is useful for the business.How do we do this?By grouping users together by some key date, we can then compare di

66900
Department of Customer Love
Looker Staff
1 year ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

Dynamic Joins - How to use liquid to dynamically adjust sql_on conditionsKnowledge Drop

Last tested: Oct 22, 2019 Define the different join conditions within liquid if/then control flow statements:explore: order_items {join: orders {relationship: many_to_onesql_on: {% if orders.id._in_query and orders.created_date._in_query %}${orders.id} = ${order_items.order_id} AND ${orders.created_raw} = ${order_items.returned_raw}{% elsif orders.id._in_query %}${orders.id} = ${order_items.order_id}{% elsif orders.created_date._in_query %}${orders.created_raw} = ${order_items.returned_raw}{% endif %};;}} In the above example, I am joining on A) the date and id fields if both the date and id field are included in the query, B) the id field if only the id field is included in the query, and C) the date field if only the date field is included in the query.Note: Make sure to confirm your if statements are in the order you would like them evaluated because the process will return the "then" response for the first true logic test.  This content is subject to limited support.               

9260
Department of Customer Love
Looker Staff
1 year ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

What is a cURL API call, and what does it look like?Knowledge Drop

Last tested: Nov 4, 2020 cURL is a command-line utility available on most Unix/Linux based operating systems which can be used to make an HTTP network request and receive the response either interactively as you type or in a bash shell script. cURL can make any kind of HTTP requests, configure request headers, and upload data in request bodies for POST, PUT, and PATCH requests. For quick-and-dirty testing or exploration of an HTTP service or REST API, curl is a simple but powerful tool.As with any use of the Looker API, the first thing you need to do when using cURL is call /login to authenticate and receive an access_token. The access_token can then be used in subsequent calls to the Looker API.For example, you would do the following (Replacing XXXXX with the appropriate credentials):curl -d "client_id=XXXXXXXXXX&client_secret=XXXXXXXXX" https://mycompany.looker.com:19999/login This will return a response that looks something like:{"access_token":"4QDkCyCtZzYgj4C2p2cj3csJH7zqS5RzK

11491
N
7 months ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

Can I have models with the same name in different Projects on the same instance?Knowledge Drop

Last tested: Sep 9, 2020No, model names must be globally unique according to docs.When Looker generates the URL for an explore, it's of the form /explore/[model_name]/[explore_name]. So each model must have a unique name on the instance in order for the URL to be assigned correctly.While it is technically possible to create LookML model files with the same name in different projects, only one of those models can be configured at a time. If you do not follow the rule of keeping model file names globally unique, you may experience strange behavior where certain parts of Looker reference the configured model and other parts reference the unconfigured model. For example, the URL of the form /explore/[model_name]/[explore_name] will always direct you to the configured model, even if you navigated there from the LookML file for the unconfigured model.  This content is subject to limited support.                  

3570
Department of Customer Love
Looker Staff
1 year ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

"Circular reference detected"Knowledge Drop

Last tested: Dec 16, 2020 Make sure you're not referencing the LookML field instead of the tables field. This example you can see that the measures sql is referencing this measure itself within the sql which causes the loop;measure: memberships {description: ""type: sumsql: ${memberships} ;;} What you should have is the ${TABLE}.field_name referencing:measure: memberships {description: ""type: sumsql: ${TABLE}.ended_memberships ;;}  This content is subject to limited support.                  

24860
Department of Customer Love
Looker Staff
1 year ago
molly.lippsett
Looker Staff
molly.lippsettLooker Staff
published in Technical Tips & Tricks

Git refresh webhookKnowledge Drop

Last tested: Jul 2, 2020 WhatThis webhook forces Looker to rescan and rebuild the referenced project, and will force changes from the filesystem to the Looker UI.https://[hostname]/webhooks/projects/[project name]/refreshWhYSome users work with LookML files entirely outside of the IDE, or copy files from one Looker instance to another directly in the filesystem. Some use machine-generated models that they distribute to their Looker instance over a script. This puts local prod ahead of remote prod. Looker doesn't constantly scan the models directory for changes - if you make a change in a directory with Looker (by editing a file, or a doing a git operation for example), then Looker automatically rebuilds the models in that directory. But if you make a change in a directory that Looker is not aware of (for instance, a new project), then Looker will not notice it and won't rebuild. This content is subject to limited support.                  

1180
molly.lippsett
Looker Staff
1 year ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

Error - "Project already Exists" for bare repoKnowledge Drop

Last tested: Jul 23, 2020 Chat Tip: Bare repos are not recommended. Use Github or another remote repository if possible.One challenge with bare repos is that deleting the project or moving to a remote repo does not delete the bare repo folder. So, if you try to use a bare repo with the same name, you'll get an error saying the repo already exists. There is no way to delete the bare repo fully from the UI, but that is a good point to add to this product feedback.If you find yourself in that unfortunate situation, and you host your own instance, follow these steps. If your Looker deployment is hosted by Looker, please open a Looker support request for assistance. Log in to the server hosting your instance. Rename ~/models/project_name to something else, like ~/models/project_name_backup. Rename ~/bare_models/project_name.git to something else, like ~/bare_models/project_name_backup.git If there once existed a remote repo with the same name, delete the remote key for that repo, found at ~

5620
Department of Customer Love
Looker Staff
1 year ago
Department of Customer Love
Looker Staff
Department of Customer LoveLooker Staff
published in Technical Tips & Tricks

Setting Up OpenID Connect (OIDC) in OktaKnowledge Drop

Last tested: Jul 21, 2020 Once logged into Okta, navigate to 'applications' tab Click 'Add Application' optionClick the 'Create New App' Ensure your new application has the 'OpenID Connect' option selected Click 'create' Name your application, and enter your Looker URI with /openidconnect added to the end of it (as described in our OIDC docs). Make sure you are NOT using the admin version of the URL! Click 'Save'Ensure you are in the 'General' tab of the new applicationScroll down to the bottom of the 'general' page and get your Client ID (which maps to the 'Identifier' field in Looker's Admin/OIDC page) and Client Secret (which maps to the 'Secret' field in Looker's Admin/OIDC page)Fill out all the necessary fields in the Looker UI as per our OIDC documentation. Take special note of which 'scopes' are required ("The OpenID Connect requires the openid scope, but your OP will likely include other scopes, such as email, profile, and groups.) As shown in our docs, the information for Auth

2470
Department of Customer Love
Looker Staff
1 year ago
molly.lippsett
Looker Staff
molly.lippsettLooker Staff
published in Technical Tips & Tricks

Troubleshooting SSO by the embed URL - First StepsKnowledge Drop

Last tested: Oct 3, 2018 This is a paraphrase of this Community article - go there for more infoUse a URL Decoder to decode the embed url and then check these things first when troubleshooting SSO by the embed URL:Does the embedded object exist on the “production” version of your instance? If it doesn't, you may see a "Model Not Found" error   Do the embed users have the correct permissions? If not, you may see a 404   Are all URL parameters are correctly formed? Troubleshooting descriptions of the parameters are in this article Doc on url parameters here Use the Embed URI Validator in the Admin panel   Confirm that everything is spelled correctly This content is subject to limited support.                  

1320
molly.lippsett
Looker Staff
1 year ago
Vincent
Looker Staff
VincentLooker Staff
published in Technical Tips & Tricks

What kinds of named value formats can I use?Knowledge Drop

Last tested: Jan 21, 2019 This Doc has the defaults.With named_value_format, you can give it any kind of excel style formatting you want. Here's the doc that shows how to set them up and some common excel style formats.Tip: At this time date formatting, color formatting, and hexadecimal conversion ARE NOT supported in Looker. This content is subject to limited support.                  

660
Vincent
Looker Staff
1 year ago
Taylor_O_Meara
Looker Staff
Taylor_O_MearaLooker Staff
published in Technical Tips & Tricks

ERROR: Measures with Looker aggregations (sum, average, min, max, list types) may not reference other measures.Knowledge Drop

Last Tested: Mar 16, 2019 The Error:You might see this error in the LookML validator or the explore.There is a related error from the same issue: "Error: aggregate function calls may not have nested aggregate or window function". What does this error mean?This error is because of a SQL limitation: nested aggregation. Most SQL dialects cannot do the following:SELECT user_id, AVG(SUM(sale_price)) FROM orders GROUP BY user_id Common Causes / Fixes:1.It's fine to have a measure that references another measure, but it must be type: number. That's because the original measure(s) that the new measure is based off is/are already doing an aggregation. Declaring a second aggregate function will cause nested aggregation. Instead, using type: number acts as a place holder so you can reference other measures or combinations of measures. Example:| measure: order_count { # Base measure #1type: countsql: ${order_id} ;;} measure: cancelled_orders { # Base measure #2type: countfilters: {field: statusval

22330
Taylor_O_Meara
Looker Staff
1 year ago
Vincent
Looker Staff
VincentLooker Staff
published in Technical Tips & Tricks

Can I use `hidden` on a join?Knowledge Drop

Last Tested: Mar 5, 2021 No, you can use hidden only at the explore and field level (documentation).BUT, there are workarounds. Method 1: use fields: [] instead on the join (documentation). This will effectively do the same thing as hidden would because it tells Looker to not bring any fields into the explore. Caveat: If any current content uses fields from this join, the content will break because the fields are no longer in the explore.For example:join: foo {fields: []} Method 2 (better way): use an empty view_label parameter in the join (documentation).For example:explore: orders {join: users {type: left_outersql_on: ${orders.user_id} = ${users.id} ;;relationship: many_to_oneview_label: ""}} Will show up as this in the explore: This content is subject to limited support.                   

14951
D
8 months ago
Vincent
Looker Staff
VincentLooker Staff
published in Technical Tips & Tricks

Could not find a set named ALL_FIELDS (all fields at the join level)Knowledge Drop

Last tested: Jan 21, 2019 That is expected, ALL_FIELDS can only be used on the explore level. This content is subject to limited support.                 

1910
Vincent
Looker Staff
1 year ago
Vincent
Looker Staff
VincentLooker Staff
published in Technical Tips & Tricks

Why wouldn’t a value_format_name or value_format apply to my field?Knowledge Drop

Last Tested: Dec 5, 2019 Without a type parameter, a field type defaults to string. You need to add type: number in most cases where this happens.You should change this:measure: ratio_of_cases_to_accounts {value_format_name: decimal_1sql: 1.0 * ${count} / ${accounts.count} ;;} to this:measure: ratio_of_cases_to_accounts {value_format_name: decimal_1type: numbersql: 1.0 * ${count} / ${accounts.count} ;;} This content is subject to limited support.                  

450
Vincent
Looker Staff
1 year ago
Vincent
Looker Staff
VincentLooker Staff
published in Technical Tips & Tricks

Can I change font size in visualizations?Knowledge Drop

Last Tested: Jul 11, 2019 It is currently not possible in all visualization styles. But this time we have font control over values in the following visualizations - Bar, Column, Scatterplot, Line, Area.We can also create a “single value visualization" out of a scatterplot which allows us to edit the displayed font size.However, we can use html tags to control font size for the time being.dimension: big_string { html: <p style="font-size:30px"> {{value}} </p> ;; type: string sql: ${TABLE}.string ;; } This content is subject to limited support.                  

37441
Rishi
6 months ago
S
Looker Staff
swhiteLooker Staff
published in Technical Tips & Tricks

"Unknown field 'X' in explore Y" or "field_filter does not support dimensions outside of the current model" for field_filter.Knowledge Drop

Last Tested: Jan 21, 2019 TL/DR: check for include: *.dashboard in other model files. This error can pop up from normal Unknown Field reasons so check those too (missing join, fields: [], etc)The issue is that the dashboard is included in a model that doesn't contain the necessary joins. Usually you have an include: *.dashboard parameter in another model than the dashboard is usually referring to.You'll either want to comment out the include: or replace the *.dashboard with specific dashboard names. This content is subject to limited support.                 

3620
S
Looker Staff
1 year ago
S
Looker Staff
swhiteLooker Staff
published in Technical Tips & Tricks

Datatype parameter for dimension groupKnowledge Drop

Last Tested: Jul 3, 2018 The datatype parameter is a parameter for a dimension_group: of type: time.This parameter tells Looker what datatype to expect coming from the database and consequentially how to transform it into the timeframes you specify in the timeframes: parameter. eg. if the column is an epoch in the database put datatype: epoch.Link to Docs This content is subject to limited support.                

340
S
Looker Staff
1 year ago
romain.ducarrouge
Looker Staff
romain.ducarrougeLooker Staff
published in Technical Tips & Tricks

IDE Quick Search/SwitchKnowledge Drop

Last tested: Jan 21, 2019 Quickly search and switch between files in a LookML Project with⌘ + J (Mac) Ctrl + J (Windows)Searches across all files in the current LookML ProjectFuzzy matching allows letters to be skipped (e.g. “ordritms” will include “order_items” as a match) This content is subject to limited support.                  

220
romain.ducarrouge
Looker Staff
1 year ago
romain.ducarrouge
Looker Staff
romain.ducarrougeLooker Staff
published in Technical Tips & Tricks

Why are my filter suggestions showing up several times?Knowledge Drop

Last Tested: May 10, 2018 The ProblemWe are seeing the suggested values for filters appearing fine in the Explore, but once a value was selected in the filter the suggestions were still listing this particular value.The ReasonThe data in the underlying table may not be cleaned. Instead of a string being "string" it might actually be " string", with whitespaces. The SolutionTry adding a TRIM() to the code for the field!sql: TRIM(${TABLE}.columnname);; This content is subject to limited support.                 

200
romain.ducarrouge
Looker Staff
1 year ago

Badges

  • Friendly Face
    sbrylowskihas earned the badge Friendly Face
  • Helpful
    carriehas earned the badge Helpful
  • Community Star
    carriehas earned the badge Community Star
  • Friendly Face
    carriehas earned the badge Friendly Face
  • Community Star
    Dawidhas earned the badge Community Star
Show all badges
Terms of Service