Sections in this category

J1QL Tips and Tricks

  • Updated

This guide contains tips and tricks that assist in creating J1QL queries in your account. If you haven't already, check out our J1QL Tutorial as well as our J1QL Language Specs articles for helpful context.

Before creating your own query, we recommend starting with our wide-array of pre-canned questions, tweaking them as needed, then creating your own from scratch.

When to Use WITH vs WHERE

WITH and WHERE are both keywords used to filter the results of your query. The filtering of the results happens either pre-traversal or post-traversal. WITH is used for pre-traversal filtering and WHERE is used for post-traversal filtering.

In a pre-traversal filter, the nodes specified in the query are filtered down before traversing child nodes and finding matching query results.

In a post-traversal filter, the entire graph is traversed to find potential matching query results before filtering occurs on the resulting graph to produce the final query result.

TIP Use WITH over WHERE when possible as the WITH operation is faster and more efficient.


  1. WITH is used to filter entity node properties pre-traversal.

Take the following example.

Find User with name~='john' 
  that HAS UserGroup
  that ASSIGNED AccessRole 
  that (ALLOWS|TRUSTS) Account


The query will find a filtered list of users with a name property that contains 'john'. The query will then find and join any connected UserGroup(s), then AccessRole(s), and finally Account(s).


  1. WHERE is used to filter entity node properties post-traversal.

Let's continue with the previous WITH example.

Find User as u1 
  that HAS UserGroup 
  that ASSIGNED AccessRole 
  that ASSIGNED User as u2 


This query will first find every User that is part of a UserGroup that is assigned an AccessRole that is assigned a User. Once the results have been been retrieved, the query will take both Users that were assigned an alias (u1 and u2), and compare the name property on each. If they match, the filtered result will be returned.

TIP Property comparisons can only be done using WHERE.

  1. WHERE is also used to filter the relationship edge properties post-traversal.

Let's look at a different example with an edge that has properties.

Find Training
  that ASSIGNED User
WHERE ASSIGNED.completedOn = undefined


This query checks the edge, ASSIGNED, and filters on its properties. In this example we find trainings assigned to Users where the completedOn date is undefined (or incomplete).

TIP Use WITH over WHERE when possible as the WITH operation is faster and more efficient.

Variable Placeholders

A variable placeholder(s) is used in the JupiterOne web application to prompt a user for a specific value(s) to be injected into a pre-saved query. Variable placeholders can be leveraged in saved J1QL queries using the double curly bracket syntax {{variable-placeholder-name}}.

Take the following example question that is saved to the question library. The query searches for an AWS instance with a specific id.


Running the above query displays a modal that prompts the user to enter a value. The query will not execute until you enter a value and press the submit button.


Create a Question that uses a Placeholder

Variable placeholder syntax cannot be used directly in the search bar. There are two ways to create a new question and add a variable placeholder.

  1. Save a New Question
  • Type a query into the search bar and enter a temporary value
  • Click the star icon to save
  • Type in a Title that describes the query (this is the question name)
  • Edit the query, replacing the temporary value with {{your variable placeholder name here}}
  • Click Save


  1. Duplicate an Existing Question
  • Click on the Library icon
  • Find an existing question with query elements similar to your new query
  • Click on the Duplicate icon
  • Update the Title
  • Update the existing query and add {{your variable placeholder name here}}
  • Click Save

j1ql-custom-query-library j1ql-custom-query-duplicate

TIP Placeholders are only supported in pre-saved queries.

TIP One or more placeholders can be used in a single saved query.

Leveraging the Graph for Context

Leveraging the graph in JupiterOne to gain context on your cyber security environment is one useful way construct a J1QL query. You can start from a basic query or use an existing one to customize a more meaningful query. Check out the J1QL Tutorial for an overview of nodes and relationships as they relate to J1QL.

Check out the Data Model - Overview document for a comprehensive list of JupiterOne entities, their properties, and the relationships between entities. The graph is a tool that can be used when these details are unknown.

When traversing the graph from a starting node, you will be able to see and expand every related node. For a comprehensive list of all entities and relationships for a specific integration, see the corresponding Integration Guide.

TIP When you are in doubt about which verb to use to traverse an edge, use the catch-all verb RELATES TO, which graphs out related entities and the relationships between them. However, you should not use RELATES TO in your final/saved query because the query will run slow having to check all nodes as possible children instead of just those connected by a specific edge verb. See the following example.

Find User that RELATES TO *

In the JupiterOne App

  1. To start, navigate to the Landing page of JupiterOne and type in a basic query, in the example below, replace DataStore with the entity class/type you are interested in.


  1. Click on the specific result you are interested in.

  2. Click on the vertical ellipses to open additional actions then click Open in Graph.


A new browser tab will be opened and will run a query that returns a graph showing everything related to or mapped to the entity via a relationship.

  1. Click on the queried entity and you will see the relationship verbs for each mapped entity.


  1. The type & class of a node can be determined by by selecting the node, clicking the info icon, and looking under the display name.


The Developer node from in the example screenshot has a type of aws_iam_role and a class of AccessRole. We know the following relationships exist for a User class.

  • A User is ASSIGNED an AccessRole
  • A User IS a Person
  • A User HAS a UserGroup
  • A User is ASSIGNED an Application

These relationships can be used to craft queries.

Find User that ASSIGNED AccessRole

Find User that IS Person

Find User that HAS UserGroup

Find User that ASSIGNED Application

TIP Relationships can be queried bidirectionally. For example, Find User that HAS UserGroup can also be queried as Find UserGroup that HAS User and will return the same results in the graph. However, the list view will return a list of User for the former or UserGroup for the latter depending on which entity you Find.

  1. Continue to expand the query by clicking on and expanding nodes in the graph. To do this, select any node and click on the ellipses.


From here we see several additional nodes. We can select a path that we are interested and continue to expand the graph until we arrive at a query that is meaningful.

In the following example we traversed from a User to a UserGroup to an AccessRole to an Account. Written out more specifically using types we have an oktauser to an oktausergroup to an awsiamrole to an awsaccount.

Our final query could be something similar to the following example, adding in filters, sorting, pagination, etc. as needed. See J1QL Language Specs to understand the query language features.

Find User 
  that HAS UserGroup 
  that ASSIGNED AccessRole 
  that TRUSTS Account



Was this article helpful?

0 out of 0 found this helpful