fix: decimal overflow in enriched packages copy

[epipav]

Note

Low Risk
Narrow change to one analytics pipe node; behavior is equivalent for valid dates and only excludes corrupt or extreme timestamps.

Overview
Fixes the ossPackages_enriched_release Tinybird pipe so the scheduled COPY into ossPackages_enriched_ds no longer fails on decimal overflow.

secondPublished no longer uses arraySort with negated toUnixTimestamp(publishedAt); it now takes the second-most-recent date via arrayReverseSort(groupArray(publishedAt)), avoiding numeric overflow from timestamp negation.

The versions filter now drops out-of-range publishedAt values (before epoch and more than one day in the future) in addition to nulls, so bad dates do not propagate into releaseCadence scoring downstream.

^{Reviewed by Cursor Bugbot for commit 726fd36. Bugbot is set up for automated code reviews on this repo. Configure here.}

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[github-actions]

⚠️ Jira Issue Key Missing

Your PR title doesn't contain a Jira issue key. Consider adding it for better traceability.

Example:

• feat: add user authentication (CM-123)
• feat: add user authentication (IN-123)

Projects:

• CM: Community Data Platform
• IN: Insights

Please add a Jira issue key to your PR title.

[Copilot]

Pull request overview

This PR fixes a decimal-overflow failure in the ossPackages_enriched Tinybird COPY pipe. The pipe computes per-package release-cadence signals from the versions datasource, where publishedAt is a Nullable(DateTime64(3)) replicated from Postgres. Out-of-range date values (far past/future) in that column caused DateTime64 (Decimal-backed) overflow during the daily COPY job, breaking the enriched dataset build.

The fix targets the ossPackages_enriched_release node by (1) replacing the arraySort(x -> -toUnixTimestamp(x), ...) descending-sort trick with the simpler arrayReverseSort(...) — removing the toUnixTimestamp conversion — and (2) bounding publishedAt to the sane range [1970-01-01, now()+1 day) so pathological dates no longer flow into the sort and downstream dateDiff calls.

Changes:

• Replace arrayElement(arraySort(x -> -toUnixTimestamp(x), groupArray(publishedAt)), 2) with arrayElement(arrayReverseSort(groupArray(publishedAt)), 2) to compute secondPublished (semantically equivalent: newest-first ordering, second element = second-newest).
• Add WHERE guards filtering publishedAt to >= 1970-01-01 and < now() + INTERVAL 1 DAY, excluding out-of-range dates that trigger the overflow.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.