Release v3.8

v3.8

A rule can be linked to a dataspot. quality indicator (DQI) straight from the rule editor, and a table can be connected by pasting the URL of its dataspot. object.
Every ruleset check then posts its measurement to the linked indicator automatically, so data stewards see live quality numbers in the place where they govern them.

Assistant replies now appear progressively, as they are written, instead of only after the complete answer is ready.

A custom SQL rule can reference virtual tables of the same connector, so a rule can join across them — for example to apply a filter from a second table. All referenced virtual tables must come from the same connector type.

The Power BI export now carries historical results per ruleset and rule over time, and keeps them even after the underlying table, ruleset or rule has been deleted — so trend and time-series reports no longer lose their past when something is removed.

A PDF-backed table can now be deleted from its own table page, instead of opening the connector’s edit dialog and removing the category there.

The in-editor coding assistant for custom SQL, Python and regex rules has now been improved with a (sandboxed) pre-execution to improve coding!

The assistant now receives an inventory of every rule already in the ruleset, including inactive ones, so it stops proposing duplicates and can recommend reactivating an existing rule instead of creating a new one.

Every issue table now shares the same column controls — column selection, a relevant-columns toggle, and defaults that persist consistently.
Charts render faster, handle single-point and identical-value series correctly, and show a proper “No data available” state instead of an empty frame.

Text rules (contains / starts with / ends with) built LIKE patterns without escaping.
A Format rule that references columns outside the ruleset’s scope now names itself and its out-of-scope columns, instead of leaving the check with an empty error.
Saving a rule with invalid code returns a proper field error again, so the Fix with AI action appears instead of a raw error message.
Custom Python rules no longer return every flagged row as their issue sample — the sample limit applies again, which previously could crash the rule editor tab on large results.
Custom SQL is validated against the rule’s sample rather than the live table.

The Improvements workflow page no longer has tabs. The issues tab always showed the ruleset’s issues regardless of what the workflow actually consumed; the source-aware preview in the Data input node replaces it.
Custom SQL rules are validated against the rule’s sample, not against the live table.

For private cloud deployments, please point your release tags to:

Platform version: 3.8.0