Designer release notes 2025

Our Designer development team works using quarterly releases. Scroll down to see the Designer quarterly releases from this year.

Note: Since the Q2 2023 release, Apteco desktop software requires a minimum of the .NET Framework 4.7.2 or later to be installed.

Note: Please see Notes for administrators for steps to be completed before applying a quarterly upgrade to your Apteco software.

Designer Q2 2025 software release

New features and improvements

The enhanced build process now supports ‘user-defined primary selectors'

A ‘user-defined primary selector’ is a Selector variable located on the master table of a FastStats system. Queries that have an 'and' clause with the primary selector run faster than other types of queries.

Support for user-defined primary selectors has been added to the Enhanced build process.

Improvements to specifying the 'People' table in core & linked systems

When building ‘Core’ and Linked systems, this feature allow you to choose a table from the linked system for licensing purposes, overriding what may have been already specified in the Core system.

Support for Palantir Foundry connections

We now support connections to Palantir Foundry using the ‘ODBC for Foundry‘ data provider.

Added ability to override Salesforce version number

The Salesforce APIs are versioned and new versions are regularly released. This new Salesforce data source functionality allows a system Designer to specify the version of the API to use when exporting data from Salesforce. Setting this version number to the latest Salesforce API version will enable the most recent data layout to be retrieved.

Additional options for default unclassified codes

Unclassifed code settings (the Code,Description and Alignment) can now be set to be the 'System Default'. This default is configured on the Advanced tab of the system configuration.

These properties are now exposed in the Define Variable property grid making it easier to change these properties for multiple variables at once.

Set KeepAlive=true for [ODP.Net](http://odp.net/) connections

The connection string parameter ‘KeepAlive=true’ is now added automatically for all Oracle [ODP.Net](http://odp.net/) connections. This setting improves extract reliability when connecting to Oracle using ODP.Net.

Warning : File format change for 2025 Q2

We’ve changed the format of the .fsDat files (full and partial) to include a header.

This change improves the FastStats service start up speed when there are a large number of partial indexes and to link the .fsDat files to the .vie file for improved safety.

Note: This change means that:
- You'll not be able to open FastStats systems built with 2025 Q2 in earlier versions of FastStats.
- You'll still be able to open FastStats systems built with earlier versions in 2025 Q2.

FastStats Designer Combine Files Pre Load Action

You now have the optional action to shift multiple incremental extract files.

This pre-load action has 2 optional sub actions:

  • Shift files

  • Combine files

The intended scenario for this enhancement is where a number of incremental extract data sources containing monthly transactions are defined.

Daily incremental extracts are appended to the end of Month01, but the other Month?? data sources are static.

At the end of the month, Month01 becomes Month02, Month02 becomes Month03, etc.

This pre-load action performs the required file moves (including the .info.xml files that hold the extract metadata), to avoid re-extraction of all the data sources.

In this example, Month04.txt then becomes Month04.bak, and Month04.bak is overwritten on future runs of the pre-load action.

This enhancement saves re-extraction of the Month?? data when it has already previously been extracted.

Bug fixes

Linked system fails to connect to core system

We’ve fixed an issue that occurred when you tried to create a new linked system. This generated an error connecting to the core system when using the linked system dialog.

Writing build log to Oracle gives a warning

Writing the build log to Oracle is not supported. We’ve found and fixed an issue which caused the following warning message to be generated when attempting to write the build log to Oracle.

Warning: Could not write Build Statistics Log to database : ORA-00911: invalid character

This warning was displayed but didn’t stop the build. If you now specify an Oracle provider for your database, then this error/warning is no longer displayed.

Input values longer than 255 characters cause missing decodes

We’ve fixed a missing decodes issue introduced in Q4 2024.

This bug prevented the auto-discovery process from finding new decodes once a value longer than 255 characters was encountered.

Process priority not overridden when run as a scheduled task

We’ve fixed an issue preventing process priority from being overridden when run as a scheduled task. Now, if the option is checked, then Designer overrides process priority and sets itself back to 'Normal' priority, with log messages to indicate this.

Virtual package doesn’t refresh all variables if folders specified in design XML

When you created a virtual package post load action, specified to ‘Refresh variables by folder’, and change this to 'All variables', Designer retained the folders setting and refreshed any variables in folders rather than all variables, as expected.

Error if key sizes are longer 255 bytes

Some of our users with unsupported key sizes > 255 bytes reported that they started to get a crash when building a system. We now generate an error message when this is the case, and limit the key size to 255 to avoid a crash.

Combine categories from database uses unclassified description as the code

Some users reported that when using the combine categories post-load action from a database the unclassified description was used as code. For example if the code was specified as ‘!’ and the description was specified as ‘Unclassified’ then the code table would be incorrect (see below), this issue has been fixed.

Extract dependency waits forever

We’ve fixed an issue that would cause Designer would wait for an extract dependency to complete even if that extract was not specified to extract. This would cause Designer to wait forever.

Combine categories UNC path issue (.dat file does not get created)

We’ve fixed an issue encountered when building a Combine Categories virtual variable with the virtual directory on a UNC path, preventing the .dat file from being created, with only the .val files being created.

Load statistics always zero for lookups when using the enhanced build process

We’ve fixed an issue encountered when using the enhanced build process, causing load statistics to always display as zero for lookups. This issue was encountered when using the enhanced data builder.

Database log credential reverts to default even if cleared

We’ve fixed an issue causing the database log credential to revert to default, even when cleared.

Although the credential was specified as default, it now stays blank, as expected.

Designer Q1 2025 software release

New features and improvements

Query support for extended text fields (up to 32,000 characters)

FastStats now supports storing and querying text data up to 32,000 characters, which is a significant increase from the previous 255-character limit. This improves data handling and also enables more comprehensive text analysis.

Key enhancements:

  • Increased text storage capacity

    FastStats text fields can now store up to 32,767 bytes.

    • For ASCII systems, this corresponds to 32,767 characters

    • For UTF8 systems, characters may consume between 1 & 4 bytes

  • Text block size increased

    To support extended text fields and improve text field compression, the text block size has been increased from 32,760 bytes to 1MB. As a result:

    • We’ve removed the configuration option for virtual text block size and compression type from the FastStats Configurator

    • We’ve applied overrides in the FastStats Service and Designer Virtual Package

    • We’ve removed the configuration option in Designer and set it to 1MB

  • Improved performance for text variable processing

    Extended text blocks improve compression ratios and reduce I/O operations, leading to better performance, particularly on file systems with high latency. Testing has shown reduced compressed text .dat file sizes and fewer I/O calls.

  • User interface and build process updates

    • AutoDiscovery has been modified to automatically detect extended text fields

    • The Standard and Enhanced Build Processes have been updated to support extended text fields

Support for DuckDB data provider and file types

This release introduces support for the DuckDB data provider in Designer. This allows you to browse tables, use the <top> placeholder, choose tables/views, and run custom queries on a DuckDB database.

In addition to the native DuckDB functionality, this update also enables support for reading the following file types directly via custom queries:

  • JSON

  • JSON.gz

  • Parquet

For more detail, see Reading Parquet and JSON files with DuckDB and DuckDB’s documentation.

Generate partial index improvements

We’ve introduced the ability to generate partial indexes in parallel, offering a significant reduction in generation time for large systems. You can now generate up to ten partial indexes simultaneously, alongside the existing sequential option.

Key updates:

  • Added UI to support generating multiple partial indexes at once

  • Up to ten partial indexes can now be generated in parallel, greatly reducing generation time on large systems

  • Partial Indexes help speed up multi-table queries and optimise system performance

Support environment variables in FastStats scripts

You can now use environment variables in the post-load action ‘Run Script’ in Designer. This makes it easier to set up duplicate systems by allowing scripts to use parameterised values rather than fixed ones.

New UInt join key type

The Enhanced Build process in FastStats Designer now supports a 32-bit UInt (unsigned integer) join key type. This is useful for systems with keys between 2.1 billion and 4.2 billion, allowing you to use efficient non-hashable join logic. This expands on the existing Int64 and GUID join key types and allows parent tables to support larger datasets.

Option to hide sensitive data in build logs

Build logs in Designer now have an option to hide sensitive data, such as environment variable values and full directory paths. This option is enabled by default for Apteco Cloud servers, improving security by preventing exposure of sensitive information.

Option to delete .dat files during the sort process

Designer now includes an option to delete .dat files during the sort phase of the build process. This can free up significant temporary disk space during the build on large systems, particularly once the merge processing has started.

However, if the build is using a delta strategy and an error occurs, then a full rebuild is required as the deleted .dat files are no longer be available.

Query engine improvements

Removal of legacy 'admin' pages from FastStats Web Service

The legacy 'admin' pages, including FileUpload.aspx, have been removed from the FastStats Web Service. This improves security and reduces maintenance of outdated functionality.

Default number of FastStats backups reduced to 3

The default number of backups for new systems in the FastStats Service has changed from 10 to 3. This reduces storage usage and minimises the need for administrators to adjust the setting when creating new systems.

Optimised distributed processing for cached cubes

Distributed processing now efficiently detects and utilises cached cube objects, improving query performance. If a cube is already cached, it will no longer be sliced, allowing the system to retrieve the cached version for faster results.

Additionally, a new currentTime parameter has been added to requests. This provides a reference time for slicing queries and cubes, reducing reliance on the todayAt parameter and improving consistency.

Faster directory listing in FastStats and PeopleStage on high-latency systems

FastStats and PeopleStage now scan files in parallel when retrieving a directory listing, significantly reducing load times. This improvement is especially beneficial for systems with a large number of files stored on high-latency network drives.

Bug fixes

  • We’ve fixed a bug that caused a crash when creating a Dynamic Virtual Variable while Global Filters are applied. Previously, this left behind invalid .val and .dat files that could cause the service to crash on start-up.

  • Running a post-load action from the command line now correctly finds and executes actions located within groups.

  • The standard build process no longer creates invalid $ftoSORTED$ files with NUL characters when Use Optimised Temporary Sort Files is unchecked.

  • We've improved the efficiency of retrieving large job results from the FastStats Web Service. Previously, the full compressed results were repeatedly retrieved and decompressed for each chunk request, causing high resource usage. Now, retrieval is optimised to reduce unnecessary processing and improve performance.

  • The Preview Categories query in Combine Categories post load actions now correctly applies environment variables, preventing errors when resolving table names.

  • Environment variables are now correctly applied to the 'Answers' query in Dynamic Database Data Sources.

  • When copying a lookup table in Designer, array field names now retain their original references and do not revert when added to the Relationships page. This prevents unintended renaming issues across copies.