Fixed an issue where mixed case column names wouldn't work as expected.

The JSON sample size is used in the Snowflake sample query

More efficient sampling of Variant data by only selecting a single column

Fixed a JSON parser bug that would cause duplication of columns in some specific scenarios

Fixed a bug that would cause some data fields to not display correctly in the data preview pane

Changing a node's materialization from a Table to a View and then deploying now works as expected.

We now correctly calculate dependencies with ref_no_link when deploying metadata, which allows deployments of graphs that are not linear.

This release includes minor improvements around Deploy. We'll return you to your regular type of release next time.

Fixed an app crash that could occur with runtime parameters in the generated documentation.

You may have noticed an error when two multi-source nodes were open simultaneously. We've fixed this.

Previously the QUALIFY keyword was not recognized by our SQL editor, this is no longer a problem.

Fixed a bug that would prevent new users from being initialized as Admins.

Sometimes Coalesce would not launch on first try — this has now been fixed.

We've tinkered and polished some rough edges. The app should be better than it was.

We’ve upgraded some dependencies that the app runs on, resulting in better performance and fewer bugs.

Sometimes workspace and environment creation dates could be inaccurate in the UI, this is now fixed.

Previously, you may have expected that specifying a Snowflake warehouse via api call would override the default value. But sadly it didn’t. It should now work as intended.

OAuth will now disconnect without making a fuss.

Generated documentation for User Defined Nodes now uses the Node Type Name like it should.

Loading the app won’t take too long anymore. Hooray!

Users would receive a message asking for excessive permissions from Azure Single Sign-On. That was bothersome, so we fixed that.

You can now actually view the Run script in generated documentation.
Occasionally, the Activity Feed did not update the run status automatically. It will now be more punctual.

The app would sometimes set multiple default storage locations. It no longer does that.

Previously, the node type editor had developed a distracting text overflow issue. That has been fixed now.

Clearing your search from the source cascader no longer throws an error.

Pagination for Job history now works as expected.

Improved issues with git versions

Improved column lineage calculation

Fixed a bug where column propagation would not create a new column correctly

Fixed performance issues with the source select dropdown

We tweaked many other issues as well

"Validate Run All" from the graph actually ran the nodes rather than appending the SQL with EXPLAIN. This is now working as expected.

Inside of the node mapping grid, we now alphabetically sort all items in the source selector.

Jinja templating errors will now be shown in the results pane to improve template development.

Various bug fixes and performance improvements.