Skip to content

Add BC performance knowledge from OptimAL learnings - #198

Merged
Jesper Schulz-Wedde (JesperSchulz) merged 3 commits into
mainfrom
jesperschulz-bcquality-coverage-review
Sep 29, 2026
Merged

Jesper Schulz-Wedde (JesperSchulz) merged 3 commits into
mainfrom
jesperschulz-bcquality-coverage-review

Conversation

@JesperSchulz

Copy link
Copy Markdown
Contributor

Summary

  • Add six focused performance knowledge articles with paired good/bad AL examples covering workload-based and overlapping indexes, buffered inserts, aggregate-before-write processing, scoped reuse, and repeated validation.
  • Refine existing guidance for partial loads, temporary-buffer costs, bulk updates, existence checks, primary-key caching, and FlowField versus stored-field totals; clarify the SIFT-specific SetCurrentKey case.
  • Extend performance-review retrieval and evaluation coverage for the new rules.

Attribution

This contribution was informed by engineering learnings and feedback gathered through the Business Central OptimAL Performance Initiative. The examples are original and generalized; no partner- or customer-specific code, identities, review results, or measurements are included.

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The package is close, but review of af94c135ad2dd7f71f8aec61916d37c2715abe79 found three substantive issues:

  1. design-covering-keys-from-read-pattern.good.al calls ByCustomerDate covering while omitting the system fields Business Central always projects (SystemId and the four audit fields). Either include every automatically projected field and discuss the write cost, or describe this as a predicate-supporting rather than covering index.
  2. cache-repeated-filtered-results-with-explicit-scope.bad.al does not establish repeated complete keys. If all item numbers are distinct, the dictionary adds overhead without reuse, yet evaluation requires a finding. Make meaningful reuse structurally evident and require that evidence in routing.
  3. The materially changed use-setloadfields-for-partial-records and prefer-modifyall-over-per-row-modify companion pairs are absent from the performance evaluation override. Register them so positive and clean retrieval behavior is covered.

I cannot formally request changes on my own PR, so this exact-head review is recorded as a comment.

Clarify predicate-supporting keys versus covering queries, demonstrate proven cache reuse, and evaluate the updated partial-load and bulk-update rules.

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Reviewed the fix at 821d62f16f90eec609a8be77ae078351ad370064. All three prior blockers are resolved: the index guidance no longer falsely claims the sample is covering once automatically projected system fields are considered; the cache fixture now establishes repeated complete-key reuse across two passes and routing requires that evidence; and the SetLoadFields/ModifyAll pairs are registered for positive and clean evaluation. All current checks pass, and I found no coupled merge-critical defect. Content is ready from this review; GitHub does not allow me to approve my own PR.

Preserve the reviewed performance package while incorporating the strengthened deterministic review-fixture coverage contract.

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Conflict resolution reviewed at b67e048070e8073f028dccbcde2cee1cec6f52b6. The merged performance routing preserves the previously reviewed guidance and incorporates current main; no evaluation registrations were lost. The strengthened changed-paired contract is satisfied by deterministic registration of the five additional changed performance articles, with retrieval cues updated where needed. Local frontmatter, knowledge index/retrieval, skill index/schema, 152 review-fixture cases, review contract, and git diff --check all pass. GitHub reports the PR mergeable and non-conflicting. Content is ready; formal self-approval is unavailable.

@JesperSchulz
Jesper Schulz-Wedde (JesperSchulz) merged commit 87ba36e into main Sep 29, 2026
7 checks passed
Michael Dieringer (MichaelDieringer) added a commit to Curabis/BCQuality that referenced this pull request Sep 29, 2026
…crosoft#198, microsoft#202) into document-distribution-batch

Resolve evaluation/review-fixtures.json semantically: data-modeling
articles list is the union of main's list and this PR's nine articles;
everything else is taken from main unchanged.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant