sxsphinxstack

Skills / Starter / Compare two database schemas

Starter Code skill

Compare two database schemas

Compare two database schemas across objects, types, defaults, constraints, indexes, permissions, functions, and migration meaning.

When to use

  • Use for migration validation, incident diagnosis, release review, or drift detection.
  • Do not apply generated destructive changes without data and rollback review.

Preconditions

  • Record engines, versions, databases, namespaces, migration revisions, and authorized metadata access.

Procedure

  1. Export normalized metadata from both sides.
  2. Compare tables, columns, types, nullability, defaults, generated values, and identity.
  3. Compare keys, checks, relationships, indexes, partitions, views, functions, triggers, and sequences.
  4. Compare roles, ownership, grants, policies, collations, extensions, and settings.
  5. Separate semantic differences from ordering or engine-generated noise.
  6. Map each material difference to migration history, data risk, and reversible action.

Failure plan

  • If engine versions differ, label unsupported equivalence rather than forcing a textual diff.

Worked example

A missing partial unique index explains duplicate active subscriptions despite matching column definitions.

Related skills

  • Compare two container manifests — Compare two container workload manifests across image identity, command, resources, security, network, storage, configuration, rollout, and metadata.
  • Create a git branch — Create a Git branch from the intended base while preserving existing changes, naming scope clearly, and verifying upstream state.
  • Compare two api responses — Compare two API responses across status, headers, schema, values, ordering, timing, and side effects.