What the Data Cleaner does
Use this browser-based page to detect and clean CSV, TSV, Excel, JSON, JSONL, XML, or YAML data. Select a file with browse files or paste content; cleanup starts automatically when the input and options are ready. The page shows the detected format, cleanup report, previews, and an export action.
Clean a file or pasted data
- Drop a file, select browse files, or paste data below. Uploaded files are detected from their extension first. Pasted text is checked for JSON, JSONL, XML, CSV/TSV, and YAML patterns.
- Choose the cleanup controls you need.
- Review the detected format, confidence, change report, and before/after preview.
- Click Export (same format) or use a routing card when one is shown.
Empty or unknown input shows an alert. Invalid input processed as JSON shows a JSON parse alert; pasted invalid JSON may instead be classified through the other detection rules.
Choose cleanup options
Select the controls that match the cleanup you need:
- Trim whitespace (cells / values / lines)
- Remove empty rows / blank lines
- Remove empty columns (tabular formats)
CSV, TSV, and Excel clean cell text and empty rows or columns. JSON trims string values recursively when enabled and always normalizes formatting to two-space indentation. JSONL removes blank lines, removes a byte-order mark, and normalizes line endings regardless of the related checkbox; trimming parsed values still follows Trim whitespace (cells / values / lines). Unparseable JSONL records remain unchanged and are reported.
XML and YAML receive text-level cleanup for trailing whitespace, line endings, the byte-order mark, and blank-line runs. The page does not repair delimiters, change schemas, or rewrite XML/YAML structure. Excel processing rebuilds worksheet cells from string values, so its output can differ in value types.
Preview and change report
The report lists detected actions and includes counts for some operations. CSV/TSV BOM and line-ending normalization may occur without a matching change-report count. CSV and TSV show table previews before and after. JSON, JSONL, XML, and YAML show truncated text blocks. Excel cleanup applies across all worksheets, while After previews only the first sheet; Before is a note. Displayed previews are truncated, but export processes the full input. A no-change result is reported explicitly.
Export the cleaned result
After reviewing the automatically cleaned result, click Export (same format). Text formats use normalized extensions: a .ndjson input exports as .jsonl, a .yml input exports as .yaml, and a .txt upload is mapped to CSV and exports as .csv. Both .xlsx and .xls input export as .xlsx.
Excel output rebuilds every worksheet from string cell values. The implementation does not establish preservation of workbook formatting or formulas. Excel processing depends on the SheetJS library being available.
When the tool routes to another workspace
CSV and TSV route to CSV & TSV Cleaner. Excel routes to CSV to Excel Converter; the route text says a dedicated Excel cleaner is coming soon. JSON and JSONL route to JSON Validator & Formatter. XML routes to XML Validator & Formatter. YAML has no linked dedicated cleaner or validator and is marked as coming soon; it receives only text-level cleanup. Email- or phone-like headers can produce domain-specific cleaning links. Advanced delimiter, quoting, schema, structural, spreadsheet, and domain cleanup is outside this tool.
Failures and input limitations
- Input processed as JSON that fails parsing shows an alert; pasted invalid JSON may be classified through another detection rule.
- Unparseable JSONL records remain unchanged and are reported.
- Uploaded text files are read as UTF-8.
- The text box display is truncated above 200,000 characters, while full text processing continues.
- No formal maximum file size, browser compatibility range, or performance limit is documented in the local implementation.
Privacy and external loading
File reading, detection, cleanup, preview, and export are implemented client-side. The page also loads analytics through shared assets, and Excel processing may dynamically request SheetJS 0.18.5 from jsDelivr. The local files do not establish a complete privacy policy or claims about uploads, accounts, or usage limits.
Frequently Asked Questions
How does format detection work?
Uploaded files are detected from their extension first. Pasted text is sniffed from its content: leading brackets and valid JSON.parse for JSON, line-by-line JSON parsing for JSONL/NDJSON, a leading angle bracket for XML, tab vs comma frequency for TSV/CSV, and key: value or list patterns for YAML as a fallback.
Will this change my data format?
CSV, TSV, JSON, JSONL, XML, and YAML use normalized output extensions. An .ndjson input exports as .jsonl, and a .yml input exports as .yaml. Both .xlsx and .xls input export as .xlsx because the workbook is rebuilt from cell values. A .txt upload is treated as CSV and exports as .csv.
Why does XML/YAML cleanup do less than CSV/JSON?
XML and YAML are indentation- and structure-sensitive in ways that aren't safe to rewrite generically. This tool only applies safe, text-level cleanup (trailing whitespace, line endings, BOM, blank-line collapsing) to those formats. XML links to the XML Validator & Formatter. YAML has no linked dedicated workspace and is marked as coming soon.
Why does it route me to another tool instead of cleaning everything here?
This shared entry point offers basic cleanup. Delimiter/quoting repair, spreadsheet-specific controls, JSON schema work, and domain cleanup such as email or phone normalization live in dedicated tools linked after format detection.