DMA-SV100-XPFv4.0.0-PV900 Release Notes

Author: Nic Napoleone, Created: 2025-06-12

Release Date: June 12, 2025

Release Overview

This release is the first General Availability release for AI Data Manipulator.

Major Features:

• Permissions

  • Segment database, section, query-level read and write access, and global settings access by user group.
  • Monitor queries inherently via an advanced SQL parser to warn or prevent users from query execution based on user permissions.

• Query Composer

  • Use T-SQL to query database tables found within your connected databases.
  • Use DuckDB to directly query tabular files found within the AI Services Engine’s Meta Filesystem storage.
  • Save queries with custom names.
  • Rename or delete your tables and files from the Query Composer page.
  • Use LLM-powered debugging to revise your queries.
  • Use LLM-powered query building to convert natural language to a SQL or DuckDB query.

• File Explorer

  • Browse the Meta Filesystem Routine and Shared directories.
  • Add nested folders to the Shared directory.
  • Save files to new locations or delete them from the directory.
  • Preview file contents.

• Audit Log

  • Track actions taken by users within AI Data Manipulator.
    • Actions List: Create, Delete, Alter, Query, Natural Language, and Permissions.
  • Select an Audit Record to view more details.
  • Export audit records within a date range to an external location.
  • Archive audit records within a date range to an external location, removing the records from the Audit Log.

Read Before Install

Due to the changes of how DMA interacts with XAT permissions, when installing make sure to Uninstall Full the current installed version in the environment. Then, install this Release Build and walk through the solution setup steps.

Dependencies

XBR-SV200-XPFv4.0.0-PV900

Xperiflow v4.0.0

Known Issues

• Database tables with non-default schema do not yet have support for in-dialog alteration.

• Extremely large datasets or query responses may fail to load due to a “Stream was too long” error. In this case, consider breaking up your datasets into smaller chunks.

• Renaming MetaFileSystem files via the Alter/Delete Dialog Flow is lacking defensive checks and may be unreliable