AI, in your work.

Three changes to evaluate separately

Microsoft lists three Business Central 2026 release wave 2 capabilities as generally available: an improved data model for loading table-extension data, indexes combining fields from a base table and its extension, and enabling or disabling indexes from AL. They appear in the version 29.0 release notes.

Nexo interpretation: these capabilities warrant reviewing queries and indexing decisions. Attributing a difference to the platform update or an extension's design requires evaluating them separately and retaining reproducible results.

What keys allow and which limits remain

The key documentation places support for combining base and extension fields in BC29, runtime 18.0. Keys still cannot include fields defined in another table extension.

The same documentation explains the index tradeoff: indexes can benefit reads, but maintaining them consumes resources during writes. Re-enabling a disabled key can require an expensive rebuild. Measure inserts, modifications and rebuild time alongside the query you intend to accelerate.

Availability and updating the environment

In Business Central online, the effective update date depends on each environment and its schedule. Microsoft's rollout process allows administrators to manage updates in the administration center. Before comparing results, check the version actually installed and prepare a test copy.

For on-premises deployments, the BC29 upgrade path supports direct upgrades from versions 25 through 28. For versions 14 through 24, it specifies an intermediate upgrade to version 25. This version path should be accompanied by an extension compatibility review.

Fictional example: querying by date and zone

As a design exercise, imagine a PedidoDemo table with a Fecha field and an extension adding ZonaNexo. The test retrieves records for a zone within a date range. The names and scenario are fictional; they represent neither a commercial extension nor a completed measurement.

First, compare identical code and data in the reference environment and BC29 without redesigning the indexes. Then, within BC29, compare one extension variant without the new mixed key and another with it. This lets you examine the effects of the upgrade and the index change separately.

• Keep data volume and distribution, filters and retrieved fields constant.
• Check that both variants return the same records.
• Define the number of repetitions in advance and record the first execution separately from subsequent runs.
• Include representative write operations and record concurrent load.

What to record for each comparison

This is Nexo's proposed record for preserving test conditions. It contains no results or improvement thresholds.

Measure the process and explain the scope

Microsoft's AL Profiler helps investigate where extensions spend execution time. Instrumentation records calls in greater detail; sampling estimates activity periodically and generally introduces less overhead. Keep the same mode when comparing profiles, and record total process duration too.

Nexo interpretation: a faster query provides evidence for that scenario. Before recommending an index change, also check writes and variability. If only one variant or set of filters improves, report that scope alongside the test conditions.