Breaking Change in List Run Endpoint – Version 7.19

Effective in version 7.19, there is a breaking change to the listRun endpoint:

• The orderBy parameter is now mandatory whenever the startingFrom parameter is included in a request. This change ensures consistent and predictable pagination behavior when using the startingFrom cursor.

What you need to do:

• If your integration uses startingFrom, make sure to include a valid orderBy value in the same request. Requests missing the required orderBy parameter will now return an error.

We are notifying all Coalesce customers using the command-line tools that an immediate upgrade to the Coalesce CLI version 7.17.1 or higher is required. Snowflake’s recent 2025_02 release includes changes that may cause service interruptions for users running older versions of the Coalesce CLI.

To prevent any disruption, please check your CLI version by running:

coa --version

If your version is lower than 7.17.1, upgrade your CLI by running:

npm upgrade -g @coalescesoftware/coa

For more information on the Snowflake bundle changes, please visit the Snowflake Release Notes.

Thank you for your immediate attention to this matter.

New

Create Node Endpoint

You can now create Nodes using the Create Workspace Node endpoint. Create Workspace Nodes supports creating all Node types. Combined with the Set Node endpoint, create and update Nodes programmatically.

Updates

• We’ve made multiple API updates including:

• Added pagination to List Environments. In addition to existing fields, we have added:

  • total - Number of results

  • startingFrom - The id the request starts from.

  • limit - The request page size. 100 is the default.

  • next - The value where the next page of results starts after.

• Added type to List Run Results. Type indicates if it is a sql or sqlTest run.

• When using ref functions, an error message will return if there is a cyclical dependency.

• We’ve made some user interface improvements including:

• Added tooltips for all statuses in the Activity Feed.

• The Fetch button in the Git Modal has moved to be in-line with commit and branches.

• Improved Column Lineage so it's easier to see the Node lineage.

Packages

New - View-Qualify Advanced Deploy Node

We’ve released a new package Node, View-Qualify Advanced Deploy UDN. This is a versatile Node that allows you to create and deploy views in Snowflake using the QUALIFY filter. It gives you a streamlined way to filter records within a View using window functions such as, RANK or ROW_NUMBER. It supports common use cases like deduplicating data, selecting top and bottom records, and applying customer filters.

Bug Fixes

• Clicking on a Node in the Docs no longer results in an error screen.

7.17.1

We’ve introduced a new feature that lets you know when a Node or Column Test has failed during a refresh.

On the Deploy page, statuses are now color coded for easy review.

• Green - It was successful with no errors.

• Yellow - The Refresh was successful, but had Column or Node test failures.

• Red - The Deploy or Refresh failed.

Take a look at Troubleshooting Deploys and Refreshes.

Power Your Documentation With AI

We’re excited to launch our Intelligent Documentation Assistant – a smarter, faster way to document your data workflows.

Maintaining clear and consistent documentation across pipelines is always a challenge. Coalesce makes it effortless by automatically generating natural language descriptions from your lineage and transformation metadata to create:

Node & Column Descriptions: Auto-generated, easy-to-understand explanations of column transformations, logic, and business context — no more manual notes or guesswork.
AI-Generated Git Commit Messages: Helpful, context-aware commit messages to keep your Git history clean, descriptive, and useful.

Benefits

How It Helps:

• Reduce Documentation Overhead: Free up time for development while Coalesce handles the docs.

• Maintain Consistency: Ensure standardized documentation across all data projects, teams, and environments.

• Improve Stakeholder Clarity: Make it easier for business users and analysts to understand transformation logic.

• Accelerate Onboarding: Help new team members ramp up faster with auto-generated, clean documentation.

How to Opt-In

AI Features not enabled by default in Coalesce. Organization admins can opt-in to the AI Documentation Assistant by going to Org Settings > Preferences and toggling AI Features.

• Column Lineage now renders correctly.

• Fixed an issue where createdAt and updatedAt fields in and older version of packages caused deployment failures.

In our ongoing efforts to enhance performance and maintainability, we're removing some legacy metadata fields that are no longer needed.

This process is happening in two stages:

• Release 7.14 - Already completed ✅

• Release 7.15 - Coming soonCompleted on March 18, 2025 ✅.

What this means for you:

• No workflows will be disrupted.

• No data will be affected.

When doing a commit, you might see changes for Nodes that you haven’t modified. This is expected when deprecated metadata is removed. You will need to:

1. Making a commit with the changes.

2. Continue on with your workflow.

Example of a commit with metadata changes

Deprecated Properties Removed:

Deprecated fields in 7.15 will affect all Source Node Types regardless of age or activity status. These fields may still appear in our API, but are no longer used in the Node metadata.

• dataset

• table

• metadata.join

• metadata.sourceMapping - already absent from most sources

See our previous changelog post about the changes in 7.14.

AI-Generated Documentation: Automatic Descriptions for Nodes, Columns, and Git Commit Messages

These features are currently in Private Preview. Please reach out to your Coalesce account manager to sign up.

Coalesce is introducing AI-generated descriptions for Node, Columns, and Git Commit messages.

By analyzing your data pipelines lineage and transformation metadata, it creates natural language descriptions. AI-generated descriptions streamlines documentation by creating clear explanations of data transformations and Git commits. Now you can focus on development while making sure your data pipelines are well documented.

Organization admins can opt-in by going to Organization Settings > Preferences and toggling AI Features.

Updates

• Run results, Column Headers now support numeric names, ensuring your data displays accurately.

• We've temporarily rolled back Workspace Onboarding based on valuable customer feedback. This change was made to ensure the best customer experience possible. We're working on improvements to deliver a refined version in the future. Sign up for our changelog email to keep to date.

Bug Fixes

• We fixed an issue where, when switching tabs in multi-source Nodes the incorrect join string was saved.

Version 7.14.3

Coalesce cleaning up deprecated metadata fields in Nodes. The change will improve data integrity across Coalesce. This will happen in two stages:

• Release 7.14

• Release 7.15

When doing a commit, you might see changes for Nodes that you haven’t modified. This is expected while deprecated metadata is removed. We suggest making a commit with the changes then continuing on with your workflow. Do not delete, edit, or modify the deprecated fields in any way.

Changes in Release 7.14

Deprecated fields in 7.14 will primarily affect older workspaces that haven't been accessed recently. For actively maintained workspaces, the impact should be minimal. However, if you attempt to commit against previously archived or long-dormant workspaces, you may encounter these changes.

Deprecated Properties Removed

The following properties will be completely removed if present:

Top-level column properties:

• ghostRecordValue

• descendedFrom

• hashColumns

• columnsForHash

Inside hashDetails object:

• hashType

• hubHashes

Properties corrected for Invalid Data

Properties set to empty string (""):

• dataType - if not a string type

• transform - if empty

Properties completely removed if invalid

• appliedColumnTests - if null

• config - if null

• hashDetails - if it's a string (should be an object)

Column Reference Fixes

Column Reference (columnReference)

• stepCounter - if null/undefined, set to the node's ID

• All IDs within column references are converted to strings (previously numbers were allowed - this change has been around for a long time)

Source Column References (sourceColumnReferences)

For each reference in the array:

• All IDs within column references are converted to strings

• stepCounter - if null/undefined, set to empty string ("")

Hashed Columns Fixes

For hashedColumns array:

• All IDs within column references are converted to strings

Upcoming Changes in Release 7.15

Deprecated fields in 7.15 will affect all Source Node Types regardless of age or activity status.

Deprecated Properties Removed:

• dataset

• table

• metadata.join

• metadata.sourceMapping - already absent from most sources