Preview Release 7.25.0
1759170230506
Updates
We’ve focused this release on UI clarity, smoother onboarding, and CLI updates.
UI Changes
Clearer Join Error When a Source Is Missing — If you attempt to join to a source Node that doesn’t exist in Coalesce, the error message now explains exactly what’s missing and where to fix it.
Connections and Authentication
Private Key Formatting Guidance — The Snowflake User Credentials screen now clarifies the expected private key format, reducing auth setup errors.
Onboarding
Skip Steps 2 and 3 — You can now skip steps 2 and 3 of the Onboarding Wizard when you already have those items configured, speeding up project setup.
CLI Updates
Readable Errors and Warnings — The Coalesce CLI now formats deployment errors and warnings for easier scanning during runs.
Bug Fixes
Mixed JSON Key Types Could Drop Derived Columns — Resolved an issue where derived mappings omitted columns when JSON keys mixed data types.
Environment Deletion via API — Fixed an error that prevented deleting Environments using the API.
CLI Hang on Exit — Addressed a rare case where the CLI could hang after completing work.
Required Metadata Discard Error — Fixed an error thrown when discarding required metadata updates.
Packages
Streams and Tasks Package
Introduced a newInsert Or Merge With Task Node type for the Streams and Tasks package.
Enables developing a stream on top of a table, view, or external table.
Adds an option to create a target table and insert or merge data from a source, with a task orchestrating the operation.
Other Updates For Streams and Tasks
Added comprehensive metadata update checks in create templates.
Fixed comment block parsing.
Updated node definition default scheduler option to prevent predecessor validation issues.
Functional Node Types
Added comprehensivemetadata update checks to the create templates flow to prevent mismatches during template generation.
Dynamic Tables
Timestamp field uses a transform when configured.
Record versioning now supports integer columns.
When multiple timestamp columns exist, only the first selected column is considered.