---
title: "Discussion on \"Building CI/CD for Oracle APEX – Part 1: Why Exports Are Not Enough\" | Hashnode"
description: "Discussion on \"Building CI/CD for Oracle APEX – Part 1: Why Exports Are Not Enough\". Building database-driven CI/CD for Oracle APEX with real DEV, TEST and PROD state comparison, PL/SQL, Oracle Autonomous Database, ChatGPT and Codex."
url: "https://hashnode.com/posts/building-ci-cd-for-oracle-apex-part-1-why-exports-are-not-enough/6ad91376d6efd7eaf6d3e0f1"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Building CI/CD for Oracle APEX – Part 1: Why Exports Are Not Enough](https://davidpataki.com/building-ci-cd-for-oracle-apex-part-1-why-exports-are-not-enough?utm_source=hashnode&utm_medium=feed)

Building database-driven CI/CD for Oracle APEX with real DEV, TEST and PROD state comparison, PL/SQL, Oracle Autonomous Database, ChatGPT and Codex.

Oct 21 · 12 min read · davidpataki.com  
Read the article

-   [#oracle-apex](https://hashnode.com/tag/oracle-apex)
-   [#oracle-database](https://hashnode.com/tag/oracle-database)
-   [#devops](https://hashnode.com/tag/devops)
-   [#ci-cd](https://hashnode.com/tag/ci-cd)
-   [#plsql](https://hashnode.com/tag/plsql)

## Written by

[David Pataki](https://hashnode.com/@enrol)

@enrol · Budaörs, Hungary

  
[View profile](https://hashnode.com/@enrol)

Oracle Database & APEX Consultant | Oracle ACE Pro

Oracle ACE Pro, founder of Enrol Consulting, and Oracle Database and APEX developer with more than two decades of experience building data-driven business applications. I write about Oracle Database, APEX, cloud and AI-assisted development.

## Responses 3

Markdown supported · ⌘↵ to post

[Sina Rezaei](https://hashnode.com/@sinarezaei)

[Oct 23](https://hashnode.com/posts/building-ci-cd-for-oracle-apex-part-1-why-exports-are-not-enough/6ad91376d6efd7eaf6d3e0f1/comment/6adb76a3faaf165abedf3156)

Dependency ordering gets tricky when a release changes both a view and the package it depends on. A diff can correctly detect both changes and still produce a broken deployment if it applies them in the wrong order. I’d want the pipeline to validate the dependency graph and test the generated deployment against a clean target schema, not just confirm that the exports look consistent.

[Julian Neagu](https://hashnode.com/@julianneagulab)

[Oct 23](https://hashnode.com/posts/building-ci-cd-for-oracle-apex-part-1-why-exports-are-not-enough/6ad91376d6efd7eaf6d3e0f1/comment/6adb02b4eae7cf0fb8b049da)

The part about hashes showing changes that don't actually matter is so real. I've run into the same thing with tools that compare generated output. Getting a clean diff is harder than it looks.

[Abhinav Sharma](https://hashnode.com/@abhinavs1920)

[Oct 21](https://hashnode.com/posts/building-ci-cd-for-oracle-apex-part-1-why-exports-are-not-enough/6ad91376d6efd7eaf6d3e0f1/comment/6ad91adfcdd62a63814fd366)

The export-as-deployment-artifact model breaks the same way in a lot of stacks. Treating current state plus an event history as the source of truth makes diffs reviewable instead of just regenerated.