Changelog
This changelog provides a detailed record of the previous 12 months of updates and enhancements we have made to improve your Dremio Cloud experience.
December 4, 2025
AI agent and AI functions now use EU endpoints for model providers when running in EU regions.
Fixed an issue where the planner threw a "cannot translate call DO_NOT_PUSHDOWN" exception.
Fixed an issue in the vectorized copier where an exception was incorrectly thrown during column vector compaction.
Queries involving evaluation of large numbers of expressions are now more responsive to cancellation.
Fixed an issue where position delete files for Apache Iceberg v2 tables were not being sorted by file_path and pos according to the Iceberg specification.
Fixed an issue with validation during the PREPARE statement from the JDBC driver for some DML queries with parameters. Querying complex types with parameters and certain DML queries selected from a subquery with parameters are not supported.
Fixed an issue where SQL queries using tables from the sys source intermittently failed with a "Failure getting source [sys]" error.
Fixed an issue of Dremio process aborting when complex data type columns were involved in hash join queries.
Fixed an issue with runtime filters on DECIMAL columns incorrectly filtering out values.
Fixed an issue where filters on VARBINARY columns could produce inconsistent results.
In the Events tab on the Engines page, the last query time now shows the correct datetime or "--" instead of an invalid date.
Added engine replica numbers to the Engines page for failed events.
Fixed an issue where deleting a namespace you were viewing resulted in a not-found error.
The Maintenance filter in the Attributes dropdown on the Jobs page now appears automatically upon entering the page.
November 20, 2025
Fixed an issue where Iceberg table write operations could create invalid Parquet files when nullability checks were disabled in Dremio, leading to DATA_READ ERROR failures during subsequent queries.
Fixed an issue where sources in a broken state could not be reconfigured.
Reflections are now blocked on JDBC sources (such as Microsoft SQL Server and Oracle) with user impersonation enabled. If user impersonation is enabled after a Reflection is created, users should delete existing Reflections to prevent unauthorized data access, as these Reflections will fail to refresh and expire eventually, but following their corresponding policy.
All values in the Attributes filter on the Jobs page now act as OR filters, showing jobs that match any of the selected criteria.
November 12, 2025
Discover, analyze, and visualize your data using natural language with Dremio's AI Agent. Use AI functions AI_GENERATE, AI_CLASSIFY, and AI_COMPLETE to integrate LLMs directly into SQL workflows and query unstructured data. Extend functionality to external agents with the Dremio-hosted MCP Server, providing secure connections to AI chat clients like Claude and ChatGPT.
The Snowflake connector is now enhanced with support for user impersonation enabling more secure and flexible access control.
Dremio now supports the Iceberg UUID data type, enabling you to work with universally unique identifiers natively in your Iceberg tables. This enhancement improves data type compatibility and simplifies UUID handling in your data pipelines.
Dremio table maintenance now utilizes file size distributions and thresholds to intelligently group and compact files, delivering dramatically faster OPTIMIZE TABLE results with much lower compute cost.
Dremio table maintenance operations are now optimized to avoid excessive computation and write operations by filtering for partitions that require optimization.
With column nullability enforcement, users can be assured on write that their data conforms to the table DDL. This simplifies the data quality experience by failing on write when something fails to match expected DDL.
Added new SET ENGINE and SET TAG commands that allow users to explicitly select an execution engine or set a routing tag for subsequent queries in a session, providing greater flexibility and optimized resource utilization for specific workloads.
Fixed an issue where columns for starred datasets would not appear in the SQL Runner's left panel if the dataset name or path contained a . character.
September 18, 2025
The Arrow Flight SQL JDBC driver now supports parameterized DML (INSERT, UPDATE, DELETE, MERGE), improving client tool integration and enhancing protection against SQL injection.
Fixed an issue that prevented adding RDBMS sources (MySQL, PostgreSQL, SQL Server, and Redshift) due to a library compatibility issue.
Fixed an UnsupportedOperationException encountered during metadata refresh of Iceberg tables in a Hive source if they did not contain table properties.
Fixed a NullPointerException that could occur when querying a table in an AWS Glue Data Catalog with an empty serdeInfo field.
Improved reliability of Amazon S3 connectivity with configurable timeout settings in milliseconds. The fs.s3a.connection.request.timeout property can be configured in the Amazon S3 connection properties or in the core-site.xml file for distributed storage.
Fixed a query performance regression related to Autonomous Reflection data collection.
Improved Autonomous Reflections with enhanced pattern detection that automatically creates more efficient Reflections to accelerate your queries.
Fixed an issue where Reflection recommendations were not generated due to inaccurate row count estimates.
Improved Reflection recommendations to generate multiple hash-matching Raw Reflections for different queries, resulting in better query acceleration.
Fixed issues with job usage info collection that caused exceptions during Reflection recommendation analysis and prevented double-counting of Arrow Flight queries.
Fixed an issue where Raw Reflections were automatically created without user changes in Advanced mode.
Added new system table function sys.reflection_refresh_settings to help administrators monitor and troubleshoot Reflection refresh configurations.
Improved Iceberg compliance for INSERT statements by writing null values for unspecified columns in Parquet files. This may impact performance proportional to the number of omitted fields.
Fixed query failures in right and outer joins with nullable fields.
The TO_NUMBER function now requires a constant, non-empty format string to improve query reliability and prevent runtime errors.
Fixed an issue where the CREATE USER SQL command would fail with an "Application error processing RPC" error.
Fixed an issue where replacing an existing view would fail if the view used the CONVERT_FROM function.
Fixed an issue where the popup that appears when an engine is starting in the SQL Runner would flicker during job execution.
Fixed an issue where opening a Reflection from semantic search would not work as expected.
August 19, 2025
Fixed an issue where queries on Iceberg tables in Hive sources could fail with an UnsupportedOperationException during metadata refresh if the tables lacked table properties.
Fixed an issue where datetime parsing functions failed with a "Can't lex input …TZH:TZM" error when using the TZH:TZM format pattern. Affected functions include TO_TIME, TO_DATE, TO_TIMESTAMP, IS_TIME, IS_TIMESTAMP, and UNIT_TIMESTAMP.
August 12, 2025
Datetime formats are now more strictly enforced, and timezone abbreviation behavior has been corrected. As a result, some queries may fail if formats are not aligned. For a complete overview, refer to Dremio's Knowledge Base.
Fixed an issue where certain queries failed with a FragmentMessageToExecutorFailed error.
Fixed an issue where a source doesn't exist error appeared for catalogs that users had access to.
Fixed an issue where Autonomous Reflections were not selecting the most optimal recommendations.
Fixed an issue where writing empty complex types to the results cache could trigger an IndexOutOfBoundsException.
Fixed an issue where queries returning zero rows with null columns were not handled correctly.
Improved Arctic catalog listing performance, reducing listing time by at least 50% in most cases.
Fixed an issue where DELETE operations could fail on certain Iceberg tables.
Fixed a NullPointerException when reading Delta tables that had undergone schema evolution.
Improved query performance by reducing initialization times and storage overhead for Reflections.
Improved read performance for Iceberg v2 tables that simultaneously contain both equality and position deletes.
Improved Amazon S3 connectivity for AWS Glue, Hive, Iceberg Catalogs, distributed storage, and Nessie sources.
Fixed an issue where Reflection recommendations would not be provided for queries involving TIMESTAMP columns.
Fixed an issue where Iceberg Data Manipulation Language (DML) queries were writing null results on some rows.
Added VACUUM TABLE support for the REMOVE ORPHAN FILES command to clean up files in Iceberg tables that are not tracked via table metadata.
July 8, 2025
Fixed an issue where the Dremio console could become unresponsive when selecting an option from the Reflection execution strategy in Advanced settings.
Fixed an issue where MongoDB queries using the $nor expression failed due to incorrect pushdown in aggregation expressions.
Fixed an issue where the Reflection refresh retry policy could fail during a Dremio upgrade.
Fixed an issue where closing a temporary tab before it could autosave would incorrectly trigger a "This script has been deleted" dialog.
June 16, 2025
Fixed an issue that caused certain queries to fail when the HASH_JOIN operator spilled to disk and triggered the exception IndexOutOfBoundsException.
June 5, 2025
Fixed an issue where metadata was not accessible for Arctic sources.
June 4, 2025
Updates to Autonomous Reflections are now shown in the history.autonomous_reflection system table.
Updated the following library to address potential security issues:
- Hadoop to Protobuf 3.15.0+ [CVE-2022-3171, CVE-2022-3509, CVE-2022-3510, CVE-2021-22570].
DX-102976
The Reflection refresh retry policy is now enforced within a 72-hour window.
Queries now scan only the required fields from complex structure type columns to prevent errors in query execution.
Fixed an issue where large case expressions were causing the planner to hang.
Fixed a memory leak on the coordinator when queries using Arrow Flight are canceled.
Fixed an issue where queries against tables with equality deletes could fail if the fields used in the deletes had capital letters.
Fixed an issue where Delta tables with unsupported minReaderVersion values were allowed to be promoted, leading to query failures.
Fixed an issue with CREATE TABLE AS SELECT statements on Parquet datasets that could result in an error.
Fixed an issue where partition recommendations were not generated for Reflections defined on views.
The following Iceberg table properties are now supported when configuring Parquet bloom filters:
write.parquet.bloom-filter-enabled.column.<col>enables bloom filters for specific columns.write.parquet.bloom-filter-max-bytessets the maximum size of the bloom filter bitset (default: 1MB).write.parquet.bloom-filter-fpp.column.<col>configures the false positive probability per column (default: 0.01).DX-100948
Fixed an issue where semantic search results did not include parent folders or source paths for entities the user had permission to read.
Queries with too many expressions in WHERE clauses or CASE statements now return a clear error message instead of failing unexpectedly.
Removed the measure_columns column from the sys.reflections table.
Running a SELECT COUNT(*) query on Iceberg tables now leverages metadata to return row counts without scanning the entire table.
Fixed an issue related to CREATE TABLE AS SELECT statements for Parquet datasets with the error message "Multiple entries with same key: $file=$file: Utf8 and $file=$file: Utf8".
Added an INCLUDE option to VACUUM CATALOG to allow targeting specific tables for vacuuming.
You can now manually refresh failed Reflections by clicking the "Refresh now" icon on the Reflections page.
The Enable asynchronous access when possible and Enable asynchronous access for Parquet datasets checkboxes have been removed from the advanced options in the source configuration dialog to simplify setup and ensure optimal performance.
May 7, 2025
Quickly discover your data using the AI-enabled semantic search. Object names, wikis, labels, column names, and other metadata are used to serve the most relevant datasets and entities.
Autonomous Reflections automatically creates and manages Reflections, accelerating query performance and optimizing data accessibility. This enables you to derive faster insights, scale effortlessly, and make data-driven decisions.
Dremio now supports a generic Iceberg REST Catalog as a source. This allows you to connect to Iceberg catalogs over the Iceberg REST API. This source is located within the Lakehouse Catalogs.
Support for vended credentials has been added to Snowflake Open Catalog, Unity Catalog, and Iceberg REST Catalog.
Dremio now supports writes against Iceberg REST sources like Snowflake Open Catalog and the generic Iceberg REST Catalog. Unity Catalog remains read only.
Results cache is now supported to accelerate repeated queries.
Incremental Reflection refresh is now supported for tables in Unity Catalog.
Fixed an issue where sources with connection issues could not be modified or deleted.
Fixed an issue where queries returning more than 1 million rows could run slower with results cache.
Fixed an issue where an OversizedAllocationException could occur for collections of variable-length values due to unnecessary data buffer reallocations.
You can now specify columns for both sorting and partitioning in Reflections when the partitioning scheme uses a transformation.
Improved Reflection refresh behavior by fixing inaccurate duration reporting and adjusting retry timing for long-running jobs and limited retry configurations.
Fixed an issue with row count estimates in Delta Lake tables that could lead to suboptimal query plans.
April 8, 2025
Updated the following library to address potential security issues:
- org.apache.parquet:parquet-avro to version 1.15.1 [CVE-2025-30065].
DX-102559
March 31, 2025
Fixed an issue where queries using Reflections failed or referenced stale materializations.
Fixed an issue where the LEAD() function returned incorrect values for the last row in a partition, where NULL was expected.
Queries accelerated by the results cache now display a lightning bolt icon and indicate a cache hit on the Job Overview page.
March 17, 2025
Dremio now supports Azure Storage container writes with the Azure Blob File System (ABFSS) driver. Iceberg metadata on Azure now defaults to the abfss:// format instead of wasbs://.
You can now use clustering as a straightforward and effective alternative to partitioning. Clustering simplifies processes and helps ensure fast queries on Apache Iceberg tables.
Dremio now supports SAP HANA as a source for AWS projects.
Azure Data Lake Storage Gen1 is no longer supported following its official retirement by Microsoft in February 2024.
You can now update Sample Source and SampleDB source settings with the necessary permissions.
Added an explicit exception for the unsupported IGNORE NULLS option in window functions.
Queries now scan only the required fields from complex structure type columns to prevent errors in query execution.
Fixed an issue where raw or aggregate Reflection recommendations were duplicated when a Reflection had no available materializations.
Fixed an issue where Reflections could have an empty Reflection score.
When writing Parquet files to Iceberg tables:
store.parquet.block-sizeanddremio.iceberg.optimize.target_file_size_mbare now ignored.- The
write.parquet.row-group-size-bytesproperty now controls row group and file size, with a default value of 128 MB instead of 256 MB. write.metadata.metrics.max-inferred-column-defaultsandwrite.metadata.metrics.defaultnow control file-level metrics in Iceberg metadata. Thewrite.metadata.metrics.max-inferred-column-defaultsproperty is now 100 instead of unlimited.- The
write.metadata.metrics.defaultmode is nowtruncate(16)instead offull. - The default value for
write.parquet.page-row-limitis now 20,000 instead of unlimited.DX-99417DX-99441DX-99743
Long project owner names are now displayed correctly.
Query completion is no longer blocked by profile updates. The profile now shows a status during updates.
February 21, 2025
Dremio supports Snowflake Open Catalog as a source.
Dremio supports Unity Catalog as a source.
February 19, 2025
Dremio supports SAP HANA as a source for Azure projects.
Fixed an issue where queries could be stuck in a pending state when executor node events are not correctly propagated to the coordinators, causing coordinators not to take appropriate corrective action.
Enhanced default raw Reflection matching to include query tree hashes so that more SQL queries can match without direct view references.
Updated the list of Dremio-supported properties and default values in the Supported Properties of Apache Iceberg Tables.
The Iceberg table property compatibility.snapshot-id-inheritance.enabled is no longer written by default when creating Iceberg tables, although it can be used for Iceberg tables (V1) for backward compatibility reasons.
Updated the following library to address potential security issues:
- AsyncHttpClient to 3.0.1 [CVE-2024-53990] DX-98578
Updated the source categories in Dremio Cloud as follows: Lakehouse Catalogs, Object Storage, and Databases.
Fixed some cases where a Reflection did not have score on the Reflections page.
Fixed an issue that could cause a query with a LIMIT clause to not be accelerated due to miscalculated row adjustments.
Fixed an issue with zero milliseconds that could occur when using the CURRENT_TIME function.
February 4, 2025
Added isPrimaryCatalog in the Catalog API, so you can update the Arctic primary source.
January 28, 2025
Fixed an issue where the snapshot ID for a table could stop refreshing and cause a Reflection to continuously refresh.
Addressed regional endpoint issues with Amazon S3 and AWS STS for users that have projects deployed in VPCs with private subnets that lack public internet connectivity.
January 23, 2025
Dremio supports Azure Synapse Analytics as a source.
Dremio supports Google BigQuery as a source.
You can now select AWS Graviton instances when creating new engines or editing existing engines.
Fixed an issue that could prevent the retry policy for Reflection refreshes from working in all cases.
Sonar cards on the Sonar Projects page now show the full name of the user who created the project instead of their email address, and when a user has been deleted from the system, you will see 
Unavailable. Users assigned to the PUBLIC role who do not have access to fetch user details no longer see a gray empty circle in place of user details.
Fixed an issue where a Reflection could have an empty Reflection score on the Reflections page.
Fixed an issue where clicking on a script in the search results on the Datasets page does not open that script in the SQL Runner.
The script lookup in the All/Mine tabs on the Scripts panel in the SQL Runner was previously case-sensitive. Now, the script lookup is case insensitive.
Because the Arrow caching feature for Reflections has been deprecated, any remaining references to Arrow cache for Reflections have been removed, such as in the sys.project.reflections system table and APIs.
You can now exclude specific tables directly in the catalog source when running the VACUUM CATALOG SQL command.
Removed unsupported pushdown operations with VARCHAR data types from PostgreSQL advanced relational pushdown (ARP) connectors.
Fixed an issue in some cases that could prevent runtime filtering if the CONVERT_FROM SQL function was used in the query.
January 15, 2025
Fixed an issue that could prevent the OpenID Connect (OIDC) option from appearing in the dropdown menu as an authentication provider for Dremio Enterprise Edition under Authentication on the Organization Settings page.
January 8, 2025
Fixed an issue that could cause Dremio Enterprise Edition to not propagate properly during the Reflection refresh process and cause SQL commands containing REFRESH REFLECTIONS to fail.