Skip to Main Content

SQLcl: MCP Server & SQL Prompt

Announcement

For appeals, questions and feedback about Oracle Forums, please email oracle-forums-moderators_us@oracle.com. Technical questions should be asked in the appropriate category. Thank you!

Migration to SQLcl Projects – Challenges, Experiences, and Best Practices

M. Tr.3 days ago — edited 3 days ago

We’re in the process of migrating our database development workflow to SQLcl projects, and while the concept looks promising, we’ve encountered some issues. I’m not sure whether these problems stem from our planned workflow or from bugs in the tool.

What We’ve Done So Far

  1. Performed a clean installation of the database schema (using our own custom tools).
  2. Used **project export** to dump the database and create an initial changeset.
  3. Installed this database on our development systems using SQLcl’s **project deploy** feature.

Problems Encountered

After deploying, if we dump the freshly installed schema again using **project export**, we see unexpected diffs caused by issues such as:

  • Inconsistent ordering (as described here)
  • Unreliable package deployment (as described here)

This is especially frustrating for hard database objects (e.g., tables) because if the object behaves the same in the database, I don’t expect any diff at all.

Why This Matters

  • This behavior is a major blocker for adopting the projects workflow (which I really like in principle).
  • For production system migrations, we plan to use Liquibase’s **sync-changelog** functionality. Therefore, it’s essential to detect schema drift accurately before migration.
  • Inconsistent diffs undermine the goal of reliable deployments and drift detection.

Questions for the Community

  • Are we using the tool incorrectly?
  • Do you have experienced similar issues?
  • Any best practices or workarounds for ensuring consistent exports and deployments?
Comments
Post Details
Added 3 days ago
2 comments
81 views