CloverDX 7.4 is here — and it's all about making it easier for different teams to work together on data.
Whether you're a business analyst building transformations in Wrangler, a data engineer orchestrating pipelines, or both, this release gives you faster feedback, less friction, and a more capable AI assistant to help you along the way.
Below is a quick look at everything that's new:
Wrangler jobs can now include formulas directly in the mapping step. Previously, adding calculated or derived values to a target layout meant inserting extra transformation steps to build temporary columns. Now you can write formulas inline at the mapping stage, simplifying job structure and reducing the number of steps needed to get to a clean output.
The mapping editor has also been redesigned to make the whole experience more comfortable. Mapping constants is easier, and you can explicitly mark a field as Not Mapped — a small but useful signal that a column hasn't been overlooked, it's just intentionally left empty.
Overall, for enterprise data teams, this means fewer transformation steps, cleaner job structure, and full visibility into what's mapped and why.
One of the more immediately noticeable improvements in 7.4 is the new dynamic preview in Wrangler. As you configure a step — before you apply it — you can now see a live preview of how your data will look as a result of that step.
The preview panel shows both input and output columns side by side, so you can focus on the columns actually affected by the step rather than scrolling through the full dataset. This is particularly useful when working with formula steps, where you want to see what a formula produces before committing to it.
Dynamic preview is also available in the target mapping editor, so you can verify your job's final output shape at a glance.
The Clover Assistant — the AI-powered helper built into Wrangler — has been significantly upgraded in 7.4. It now understands formulas far better than before, which means it can generate more complex formula suggestions and help you work through formula-based mappings more effectively.
Two improvements stand out in particular:
Syntax error fixes: The Assistant can now suggest fixes for syntax errors in formulas — catch a typo or malformed expression and get a corrected version without hunting through documentation.
Logical error fixes: Beyond syntax, the Assistant can also identify and suggest corrections for logical errors in formulas — cases where the formula runs but produces the wrong result.
Combined with the new formula-aware mapping, this improvement makes the Clover Assistant a useful co-pilot for building and debugging Wrangler jobs, not just a way to get started quickly.
A common pattern in CloverDX deployments is having domain experts build data mappings in Wrangler while engineers handle the broader pipeline orchestration in Designer graphs. Until now, connecting these two worlds required exporting a Wrangler job and converting it to a subgraph — an extra step that created maintenance overhead whenever the Wrangler job changed.
The new WranglerJob component removes that friction entirely. You can now reference a Wrangler job directly from within a graph and run it as a subgraph with no export or conversion required. The component picks up the job definition automatically, keeping the two in sync.
You also get additional control from the graph side: override data source parameters, or bypass the source or target configured in the Wrangler job entirely if your pipeline requires it.
Domain experts own the mapping logic; engineers own the pipeline. CloverDX 7.4 makes both sides easier without requiring either to change how they work.
Data Apps and Data Manager are frequently used together — a Data App might handle loading data into a set, running ad hoc validation, or triggering a downstream process. In earlier versions, launching a Data App from within Data Manager required navigating away or knowing where to find the app.
In CloverDX version 7.4, you can link Data Apps directly to a data set in Data Manager. They appear as icons in the toolbar, so users can trigger them with a single click from the context they're already in. When the app completes, a notification gives users direct access to the output — whether that's a downloaded file, a results summary, or something else.
Alongside the Data Apps integration, 7.4 includes a number of improvements to Data Manager that give both users and administrators more control:
Granular row permissions: New row-level permissions let you control whether rows can be inserted or deleted in a given data set, independently of each other. Useful for data sets where the structure should be fixed or where additions need to go through a controlled process.
Better editing for long values: An improved editor for long and multi-line values makes editing complex content in data sets significantly easier.
UX polish: Additional context menu items, an improved column chooser, and settings screen refinements across the board.
CloverDX 7.4 expands the list of officially supported deployment configurations to include Microsoft Windows Server 2025 Standard.
CloverDX Designer now requires Java 21 (the Eclipse platform has dropped support for Java 17), though CloverDX Server continues to support both Java 17 and Java 21. If you're running Designer on Java 21 and targeting a Server on Java 17, set the Designer's Java compatibility level to 17 to ensure any Java classes in your solution compile correctly.
Multiple output ports on Merge and Concatenate: Merge and Concatenate components now support any number of output ports, matching the behaviour of components like SimpleGather and reducing the need for SimpleCopy in some graph layouts.
Formatted descriptions: Schedules and event listeners now support formatted descriptions using an HTML editor, making it easier to highlight important information at a glance.
Proper HTTP 429 responses for rate-limited Data Services: Data Services now return HTTP 429 (Too Many Requests) when rate limits are exceeded, along with new headers that let callers understand their current limit and remaining capacity.