Werkzeug · CSV viewers and editors
Side-by-side CSV compare built for privacy-conscious reviews
Use Table to load two CSV files side by side in the browser. When headers match, see differing cell counts, jump to changed rows only, and keep both extracts on your device.
Operations and data teams rarely need a heavyweight ETL job when the question is simply whether two flat files diverge. They need fast optics: row counts, column parity, and a tight list of rows that changed between exports.
Table’s compare experience mirrors the main grid: familiar sorting, filtering, and search, but read-only so reviewers do not accidentally overwrite a golden extract while chasing deltas.
Because both files parse entirely in the browser, you can run the workflow on VPN-protected laptops or air-gapped-adjacent setups where uploading a second CSV to a random “merge” website is off limits.
How the two-panel layout helps reviewers
Left and right pickers keep responsibilities obvious: staging on one side, production on the other, or this week versus last week. After both files load, a summary explains whether column structures align and quantifies how many rows and cells diverge.
When structures match, a checkbox trims the grids to rows with any difference so you are not scrolling thousands of identical lines. Each panel keeps its own pagination and filters, which helps when one side is wider but you only care about overlapping keys.
Limits, edge cases, and when to pre-process
Compare mode inherits the same import ceilings as the editor to protect browser memory. If you bump the cap, pre-aggregate or segment files before loading. Truncated imports still compare what loaded, but totals should be interpreted against the trim warning.
Exports with different quoting, delimiters, or header labels may fail the automatic diff gate even when business meaning matches. Standardize exports at the source, or normalize headers in a trusted pipeline, then re-run compare for meaningful stats.
Pairing compare with export and QA workflows
After you validate deltas, download cleaned subsets from the main viewer if you need editable grids again. Compare mode is intentionally read-only so golden files stay immutable while you investigate.
For recurring checks, nightly inventory, hourly order pulls, script ordering upstream so row indexes stay comparable, then spot-check visually in the browser before you promote automation.
Häufige Fragen
- Is this the same engine as the main Table viewer?
- Yes. Parsing, column typing, and grid rendering share the same stack. Compare mode disables edits and paste so you can focus on spotting differences safely.
- Can I compare Excel files directly?
- Upload CSV exports. Save spreadsheets as CSV from Excel or Sheets first so headers and delimiters are consistent with what the parser expects.
- What if I only care about one column?
- Load both files, confirm columns match, then use filters or search within each read-only grid to isolate the field you care about before toggling the differences-only view.
