USER MANUALS

What Is New in Denodo 9.4

This page lists the main enhancements of the update 9.4 of the Denodo Platform and the Solution Manager.

We also published these videos that explain the enhancements included in this update:

For the full list of enhancements and bug fixes, read the Release Notes in the Support Site.

Virtual DataPort

  • New data type: vector. This type is designed specifically for AI and Machine Learning workloads. Conceptually, a vector is an array of floating-point numbers (e.g., [0.12, -0.5, 0.88, …]). Vectors represent Embeddings. An embedding is a numerical representation of unstructured data - like a sentence, an image, or an audio file - captured by an AI model. While you could technically store these numbers in an array of float values, this new data type is specialized for distance metrics; that is, to calculate how “similar” two vectors are. To do this, Virtual DataPort now has several functions to calculate this: VECTOR_L1_DISTANCE, VECTOR_COSINE_DISTANCE, and more. See all these functions at Embedding Functions.

    As with other data types, Virtual DataPort can read these values from databases, cache values of this type, etc.

  • The Denodo Assistant is a set of features of the Denodo Platform that use artificial intelligence to make it easier to create and use virtual models. This update includes more features:

    • In the panel Denodo Assistant of the VQL Shell:

      • Explore metadata: you can now make queries about the metadata of the views.

      • Users can add their own “custom instructions” to provide additional context to the Denodo Assistant requests.

      • Users can limit the databases they want to work with.

    • You can now configure the Denodo Assistant to use these vector databases:

      • Amazon Aurora PostgreSQL

      • Azure SQL Managed Instance

      • GCP Cloud SQL for PostgreSQL

      • Microsoft SQL Server 2025

      • Oracle 23ai

      In 9.3, PostgreSQL (PGVector) was already available.

Virtual DataPort - Data Sources

  • New JDBC-ODBC bridge developed by Denodo. It enables connections to databases that support the ODBC API but lack native JDBC support.

    Virtual DataPort already had a JDBC-ODBC bridge for this purpose, but it had been deprecated since Denodo 8.0 because most databases provide a JDBC driver. The old JDBC-ODBC bridge also had issues when running on some architectures and operating systems. Due to this, this update includes a new JDBC-ODBC bridge.

    In Design Studio, consider this:

    • To connect to Microsoft Access, use the new adapter labeled Microsoft Access.

    • In the wizard to create data sources, the connector Generic (ODBC) uses this new JDBC-ODBC bridge.

    • The old data source is now labeled as Generic (Legacy ODBC).

  • Database adapters:

    • New SQL database adapters:

      • Amazon Athena 3.x

      • Apache Doris (also as cache database)

      • Microsoft SQL Server 2025 (also as cache database)

      • Oracle 23ai

      • Adapters built over the new JDBC-ODBC bridge:

        • Microsoft Access: this one replaces the existing Access adapter.

          To use it, you must create a new data source (i.e. in the wizard to create data sources, select Microsoft Access).

        • SAP CDS (ABAP)

        For these ones, before creating the data source in Design Studio, you must create a Data Source Name (DSN) in the computer in which the Virtual DataPort server runs.

    • New no-SQL adapter: Amazon DocumentDB.

    • Snowflake: added support for key-pair authentication. You can store the key-pair file in the data source itself or configure the data source to obtain it from the credentials vault of your organization.

  • JSON and XML data sources with HTTP Data route: you can now configure them to return the HTTP response code of the requests as a field of the base views (look for the option Include the HTTP response code in the output in the configuration of the data source).

  • Data sources with HTTP Data route (DF, Excel, JSON, XML and Custom): you can now configure these data sources to obtain the credentials from the credentials vault of your organization.

Virtual DataPort - Execution Engine

  • Several improvements in performance. For example, improved optimization for queries that use function expressions within the GROUP BY clause. This enhancement allows for more efficient query planning, specifically by improving the ability to push down the GROUP BY operation through JOIN and UNION operations.

Virtual DataPort - Materialization Engine (previously known as Cache Engine)

When enabling materialization full in a view, the tables that will hold the data are now created without “control columns”. This becomes the default behavior and changes how the data is materialized and invalidated. It speeds up the process of materializing data. For views which cache was enabled before this update, the data is materialized the same way as before (see Cache Tables Without Control Columns).

Virtual DataPort - Version Control System (VCS)

The administrator can now restrict which users can merge changes into the main workspace. To enable this, open the VCS Settings and switch on Protect main workspace. After enabling this, only the users with the role integrate_workspace can merge changes into the main workspace.

Virtual DataPort - Design Studio

In the wizard for creating derived views, the tab Output has two new features that make it easier to transform base views obtained from a REST API:

  • Project multiple registers at once: users can now project several registers simultaneously.

  • Subfield prefixing: when projecting the subfields of a register, an optional prefix can be added to the name of the resulting subfields.

Virtual DataPort - JDBC Driver

The Denodo JDBC driver now includes support for two-factor authentication (2FA) (more details).

Virtual DataPort - ODBC Driver

The Denodo ODBC driver for Windows is now digitally signed. This prevents Windows from blocking the installation.

Data Marketplace

New Denodo Assistant Chatbot. This is a conversational interface designed to transform how users interact with the Data Marketplace. By combining the capabilities of Large Language Models (LLMs) with a metadata-enriched Vector Database, the chatbot allows users to discover data, explore views, and understand complex relationships between them using natural language (see Denodo Assistant Chatbot).

  • We now provide the language file for Data Marketplace in Italian. The page Language of Data Marketplace, Design Studio and Scheduler explains how to deploy it. This page also explains how to create your own language file for these applications in case the one you are looking for is not listed there.

Lakehouse Accelerator

Scheduler

  • Added support for JDBC condition-based triggers to Scheduler jobs (more details).

  • Scheduler now supports Elasticsearch 8.x and 9.x (see Elasticsearch Sources).

  • Added support to create JDBC data sources to:

    • Oracle 23ai

    • Microsoft SQL Server 2025

    • PostgreSQL 14, 15, 16, and 17

    See the full list in JDBC Drivers.

  • You can now configure Scheduler to store the metadata in Google Cloud SQL for MySQL and GCP Cloud SQL for PostgreSQL.

Solution Manager

Solution Manager can now deploy revisions concurrently to all the servers of an environment, if they use Derby as their metadata database.

Enhancements that Apply to Several Modules

Accessibility enhancements in the Data Marketplace, to ensure compliance with WCAG Level AA standards for assistive technologies. Additionally, accessibility improvements were also made in Design Studio, Scheduler, and Solution Manager.

Information about Deprecated Features

With the release of this update, we have deprecated these features:

  • In Scheduler, the support for Elasticsearch 7.x is deprecated. You should upgrade to Elasticsearch 8.x or 9.x.

  • In Virtual DataPort, the support for the adapter Amazon Athena 2.x is deprecated. You should use the new adapter Amazon Athena 3.x.

The page Features Deprecated in Denodo Platform lists all the features that are deprecated.

Add feedback