Skip to main content

Release Notes

For a detailed list of all changes, see our Changelog.
For migration tips, see our Migration Notes.
For details about versioning, see our Versioning Policy

Dependencies​

Atoti server 6.1 requires :

  • Java 21
  • Spring 6
  • Spring boot 3.2

6.1.5​

2025-03-25

🪄Improvements🪄​

📱Datastore​

  • Add a new method to rebuild all indexes of a store
    Rebuilding an index is a maintenance task that can help reclaim memory.
    It is now possible to rebuild all indexes of a store using IDatastore.rebuildIndexes(String storeName).

6.1.4​

2025-02-27

✨New Features✨​

🔎Observability​

  • Add a new diagnostics endpoint and a measure lineage service [doc]
    This service allows you to retrieve the lineage of your cube measures.
    A default user interface will soon be available in the Admin UI.

🪄Improvements🪄​

📤WebServices​

  • Add support for distinct queries on the Database REST endpoint [doc]
    This improvement is available on the new version of the Database API /activeviam/database/v10/database.
    This is the only difference between v9 and v10 of the REST endpoint.

🔎Observability​

  • Add List/Distinct/GetByKey datastore queries tracing [doc]
    Query tracing from MDX to Database retrieval is now fully complete.

🤖AI-based query optimizer​

  • Improve optimizer performance [doc]
    The AI-based query optimizer now takes advantage of the support for multi-column distinct count approximations in the database API. In versions prior to 6.1.3, after having generated a recommendation of partial aggregate provider, the optimizer would run a query against the contributors.COUNT of the associated cube. The result of this query represents the number of locations/lines stored in the provider, and could thus be correlated to the memory footprint of this provider.

New Features​

Improvements​

Starting from version 6.1.4, the optimizer will instead run a statistics query against the underlying database, thus running much faster, while being less costly in the case of an external database.

6.1.3​

2025-01-13

Improvements​

Datastore​

  • The Datastore now supports multi-column distinct count approximations.
    This new functionality has been used to:

    • Increase the precision of the single column statistics query. Its previous implementation would always provide an overestimate.
    • Allow to estimate the cardinality of a combination of several columns

    The entry point for these functionalities is IQueryRunner.statistics().estimateColumnCardinality(...).

Cube​

  • Expand usability for virtual hierarchies [doc]
    They are now usable from an Excel client and within a horizontal distributed context.

Aggregation engine​

  • Improve measure computation default partitioning in DirectQuery context
    When no partitioning is implied, default partition is now chosen based on query filters and level cardinalities.

Observability​

  • Improve query tracing [doc]
    MDX query journey is now entirely traced from endpoint call to database retrieval. Lookup queries are not traced in the Datastore setup.

Known issues​

Aggregation engine​

  • Issue with the Copper API behavior in a distributed setup
    Since version 6.1.0, the Copper API has been available for use in a distributed setup.
    Unfortunately, an undetected bug might lead to inaccurate measure computations in this limited setup.
    In some exceptional conditions, data nodes measure graphs could be de-synchronized.

6.1.2​

2024-11-28

New Features​

DirectQuery​
  • New distributed date roll over procedure with downtime [doc]
    This procedure describes how you can roll a date between data nodes with a minimal downtime.

Aggregation engine​

  • New MaxAppendOnlyFunction and MinAppendOnlyFunction native aggregation function
    These functions does not keep a history and thus does not support fact disaggregation.
    But, they can be used in aggregate providers.

Improvements​

Copper​

  • Define continuous query handler for your copper measures [doc]

Observability​

  • New available metrics [doc]

Changes​

Spring​

  • Atoti server beans name are now qualified with AtotiServer- to avoid collision

Fixes​

DirectQuery​

  • Fix aggregate table bootstrapper on specific cases
    Multiple hierarchies on the same selection field. Aggregate provider with native measures.

6.1.1​

2024-10-17

New Features​

AI Query Optimizer​

  • AI-based aggregate provider recommendation (experimental) [doc]
    Optimizer can analyze your query patterns and recommend an aggregate provider configuration.

Live extension​

  • Extend your java application on the fly with Atoti Python API [doc]
    Expose your java application with the runtime starter.
    Then use Atoti Python API to customize your application at runtime.

Improvements​

Sources​

  • Cloud multiple files CSV topic
    Load multiple CSV files in a topic from the cloud with CloudMultipleEntitiesCsvTopic.

Distribution​

  • Ensure all your measures are serializable with activeviam.activepivot.ensureSerializableMeasures
    In a distributed context, all your post processors measures should be serializable. This allows their transmission to the query node.
    You can activate this property to check, at startup, if your measures are serializable.

Fixes​

DirectQuery​

  • Fix time travel activation Snowflake and Databricks
    From 6.1.0, it was not possible to activate time travel for Snowflake and Databricks.

Aggregation engine​

  • Fix default partitioning in DirectQuery context
    Before this fix, in a DirectQuery context, when no aggregate provider was defined, measure computation was not partitioned.
    Now, Aggregation engine applies a default partitioning.
    Definition of a partitioning on the jit is now forbidden.
  • Fix timeout propagation for drillthrough queries
    Before this fix, the timeout was not propagated to the database on drillthrough query.
    Now, the timeout is correctly propagated.

WebServices​

  • Fix incorrect empty location filtering in data export
    Before this fix, tuples with no data was filtered from export.
    Now, they are correctly exported.

Deprecations​

  • Source API streamlining deprecations
    A source API rework is planned for 6.2.0.
    Preparing this work, several API have been deprecated (with alternatives).
    Refer to the change log for an exhaustive list of deprecations.

6.1.0​

2024-09-06

Public and Internal APIs​

In this new release, the boundaries of public and non-public classes/APIs have been clearly delimited.

There are now two broad types of classes for use in Atoti:

  • Public apis, marked with .api.. These classes can be used in any project. The public API are guaranteed to remain compatible between bug fix releases. Moving to the latest bug fix will in time become frictionless.
  • Not public, marked with .private., .internal. or .avinternal are not destined for use outside ActiveViam’s own developer teams. These classes have no guarantee to be compatible between bug fix releases.

A tool for automatically reorganizing imports is available in the migration notes.

This major change aims to simplify maintenance for both ActiveViam and customers.

Projects that still utilize classes that reside in a package that is not part of the delimited public API after migration to this new version should contact ActiveViam's support teams to find a suitable solution.

Discussing the technical challenges behind the use of an internal API will shed light on how Atoti Server is used, what it is used for, and will help shape the future of the product's API.

New Features​

Distribution​

  • Copper in distributed cubes is now stable [doc]
    Copper calculations can now be defined in Query cubes.

DirectQuery​

  • DQ API V2 [doc]
    Full control on table and field building.
    Common runtime interface for all databases.
  • Emulated time travel (experimental) [doc]
    Direct query can now leverage defined versioning columns in tables to provide a versioned experience on external databases which does not supports time travel.
  • Official support of Dremio database [doc]
  • Aggregate table is now stable [doc]
  • Incremental refresh is now stable [doc]

Project building​

  • Spring Boot Starters [doc]
    Spring Boot Starters are now available to help with building an Atoti Server application.
    As of now, the following modules can be used in your projects:

    • com.activeviam.springboot:atoti-server-starter
    • com.activeviam.springboot:atoti-ui-starter
    • com.activeviam.springboot:atoti-admin-ui-starter
    • com.activeviam.springboot:atoti-server-apm-starter
    • com.activeviam.springboot:excel-addin-starter

    See the Sandbox for an example of their usage in a project.

Data structure​

  • SameValueVector:
    Newcomer in vector computation components allowing to manipulate a same value vector.

Improvements​

DirectQuery​

  • Multithreaded result set parsing
  • Queries with inequality operators (<, <=, > and >=) can now use aggregate table
  • Clustering fields can be used by runtime queries (not only feeding)

Copper​

  • Switch measure (experimental) [doc]
  • Database measure (experimental) [doc]

Mdx Engine​

  • Calculated measures evaluation [doc]
    The Mdx engine will now analyze the expression of each calculated member in order to find a better companion than contributors.COUNT. This will reduce the computation time of some calculated measures. Also, this optimization aims to improve polymorphic cluster performance by reducing the number of queried data nodes in some cases.

Distribution​

  • Option to conceal branches from the distributed application
    The query node will have no knowledge of some of all the branches created on the data nodes. One can also change if a branch is concealed or not on the fly.
  • Execution of PostProcessors in horizontal distribution [doc]
    PostProcessors will now be executed in data nodes only if they opt-in for.

Test API​

  • Public API for testing [doc]
    Atoti Server now ships an official API to test your Atoti Server project. This API is available in the module com.activeviam:atoti-server-test. It provides a series of helper classes to run queries on ActivePivot instances and check the results as well as a helper class to write tests involving the Distribution layer.

Observability​

  • Observability is now implemented with OpenTelemetry [doc]