CSV Viewer

CSV Viewer is a client-side table preview for comma-separated values. It parses the first row as headers, renders a sortable grid, and keeps the full text in memory so you can inspect exports before ETL, JSON conversion, or ticket attachment.

What you get

• File picker / drag-and-drop load via FileReader (text)
• RFC 4180-style parsing: commas, quoted fields, escaped quotes
• First row treated as column headers
• Clickable header sort (localeCompare, ascending/descending)
• Row and column counts in the table chrome
• Local-only processing — spreadsheet contents are not uploaded

Use this viewer when

• You need a quick look at an export without opening Excel
• You want to verify headers and row shape before JSON/CSV transforms
• You are checking whether quotes and commas survived an export
• You must stay offline with customer or payroll CSVs

Load → scan headers → sort a column → decide handoff or fix encoding.

1. Open a .csv from disk (or paste-compatible workflow by loading a saved export).
2. Confirm the header row names match the producer’s schema.
3. Check row count against the expected export size.
4. Click a column header to sort; click again to reverse.
5. Scan for ragged rows (empty trailing cells) or merged-looking values.
6. If characters look wrong (é, ’), fix encoding at the source and reload.
7. For transforms, copy the CSV into JSON ↔ CSV, Dev Toolkit Columns → JSON, or your ETL path.

The parser is comma-delimited. The first non-empty logical row becomes headers; every following row is a data row aligned by column index.

What is detected

• Comma (,) as the field separator
• Double-quoted fields that may contain commas and newlines
• Escaped quotes as doubled quotes ("") inside quoted fields
• CRLF and LF row terminators

What is not auto-detected

• Semicolon (;) or tab (\t) delimiters common in EU Excel exports
• Separate “has header” toggle — row 0 is always headers
• Type inference (numbers/dates stay strings in the grid)

Practical tip

If your file uses semicolons, convert to commas (or re-export as CSV UTF-8) before opening here. Mis-delimiter files often appear as a single column.

Sorting is lexicographic via localeCompare on cell strings. It is ideal for scanning categories and IDs, not for numeric magnitude order.

• Click a header to sort ascending; click again for descending.
• Numeric columns sort as text (“10” before “2”) — expected for a preview grid.
• Empty cells sort as empty strings; ragged rows may shift values visually left/right.
• Sort state resets when you load a new file.
• Use the row × column meta line to sanity-check export completeness.

A vendor sent overnight.csv. Before you load it into a warehouse job, you need to confirm headers, row count, and that quoted commas did not split fields.

How CSV Viewer solves it

1. Open the file and read the header row aloud against the contract schema.
2. Compare displayed row count to the vendor’s manifest.
3. Sort a key column (order_id, email) and spot blanks or duplicates visually.
4. Open a suspicious row’s cells — quoted address fields should stay in one column.
5. Reject or re-request the file if the shape is wrong before ETL runs.

You catch schema drift in seconds instead of after a failed pipeline run.

You need objects for an API mock or LLM tool schema, but the source of truth is still a CSV export.

How CSV Viewer solves it

1. Preview in CSV Viewer to confirm headers are clean (no blank names, no duplicate titles).
2. Copy the CSV text into Dev Toolkit → Columns → JSON for an array of objects, or use JSON ↔ CSV tools for round-trips.
3. Validate a sample object in JSON Formatter before wiring mocks.
4. Keep the CSV as the audit artifact; commit only sanitized samples.

Headers verified in the table become predictable JSON keys for downstream tools.

Names show as José or smart quotes appear as ’. Stakeholders think the data is corrupt.

How CSV Viewer surfaces it

1. Load the file — FileReader uses the browser’s default text decoding (typically UTF-8).
2. If mojibake appears, the file is likely Windows-1252/Latin-1 mislabeled as UTF-8 (or the reverse).
3. Re-export from Excel/Sheets as CSV UTF-8, or convert with a proper encoding tool.
4. Reload and confirm diacritics and currency symbols render correctly.

You identify encoding mismatch at preview time instead of after Unicode-breaking imports.

The entire file is held as a string and parsed into a full in-memory table. Browser tab memory — not a streaming engine — is the ceiling.

• Multi-hundred-MB CSVs may freeze or crash the tab — split files or preview a head sample first
• Sorting copies the row array; wide + tall tables amplify memory use
• There is no virtualized infinite scroll for million-row sets — use a database or CLI (csvkit, DuckDB) for those
• For spot checks, export the first 1–5k rows from the source system into this viewer

Symptoms: some rows show fewer cells, values appear under the wrong header, or a field that contains commas splits into extra columns.

Why it happens

Unquoted commas inside fields, mismatched quotes, or rows truncated mid-export break column alignment. Semicolon-delimited files parsed as commas collapse to one column.

Diagnose

Find the first bad row. Check whether address/name fields contain commas without surrounding quotes. Confirm the file uses commas, not semicolons.

Fixes

1. Re-export with proper quoting (Excel: CSV UTF-8; databases: FORCE QUOTE).
2. Convert semicolon/TSV to comma-separated before opening.
3. Fix unbalanced quotes in the offending row.
4. Reload and verify column count matches the header length on sample rows.

Symptoms: accented characters, symbols, or CJK text render as Latin junk (Ã, Â, â€).

Why it happens

Bytes were decoded with the wrong character set. This viewer reads files as text via the browser (UTF-8 expectation for modern CSV).

Diagnose

Open the same file in an editor that shows encoding. If it looks correct as Windows-1252 but wrong here, convert to UTF-8.

Fixes

1. Re-save as CSV UTF-8 from the spreadsheet or BI tool.
2. Convert with iconv or a dedicated encoding converter, then reload.
3. Ensure producers declare UTF-8 in pipelines (especially legacy ERP exports).

Fix encoding at the source — sorting and JSON handoff will only propagate the corruption.

CSV exports often contain PII, payroll, or customer lists.

• Prefer local preview over uploading to unknown “CSV online” sites
• Strip or hash emails/IDs before sharing screenshots of the grid
• Do not commit production CSVs to public repos
• Watch for formula injection (=, +, -) if pasting into spreadsheets later

• Agree on UTF-8 + comma + quoted headers with every data producer
• Preview headers before any automated JSON conversion
• Use a head sample for huge extracts; reserve this tool for inspection, not warehousing
• Document column meanings next to the file — headers alone are rarely enough
• After fixing delimiter/encoding, re-verify row counts before ETL