CSV Reconciliation for Migration

Use this checklist when cutting over data between systems.

When to use this

  • Before flipping reads to a new database.
  • After an ETL run that should be lossless.
  • When the row count from source doesn't match the target.

How this checklist works

Some items run a quick local check (e.g. scanning sample responses for secrets or PII). Others are manual confirmations or open another tool. When you finish, copy the report as Markdown to paste into your PR description, ticket, or Slack.

Privacy

This page is a static HTML file. The checklist component runs locally in your browser. The raw text you paste into the Run boxes stays in memory and is never saved. Your progress — checked items, notes, owner/reviewer, and each check's verdict and trimmed output — is saved to localStorage in this browser so you can resume, and you can remove it any time with Reset or the footer's "Clear local data" action.

Related tools

Other release checklists

What to capture during this CSV migration

Use the checklist as a release note template, not just a pass/fail list. Capture source row count, target row count, join key, duplicate-key report, encoding assumptions, and a sample of mismatched rows. This gives reviewers enough context to understand what changed and what still needs follow-up.

Common blockers

  • Evidence is copied from staging but the production path, host, or namespace is different.
  • A secret, token, cookie, or personal field appears in an example that will be shared outside the team.
  • The release owner signs off without a clear rollback note or known-risk comment.

Helpful next checks