CSV to Log Converter

CSV is the universal exchange format for log analysis — every BI tool, auditor and analytics export hands you a table. But when you need to send that data back into a log pipeline, replay events into a test harness, or simply read rows in a terminal, a table of columns is not a log. CSV to Log Converter reverses the pipeline: it takes CSV rows and renders each one back into a greppable text line, entirely in your browser.

The parser handles the real-world mess of CSV: comma, semicolon or tab separators, quoted cells with embedded separators, embedded quotes and embedded newlines inside a single cell, and both header and headerless layouts. With a header row, columns are read by name; without one, you flip off the header switch and either let columns default to col1, col2… or type your own comma-separated labels. Rows with ragged column counts (a trailing field missing here and there, common in hand-edited exports) don’t break the parser — missing cells simply come through as empty strings.

You map which columns carry the timestamp, level, category and message using dropdowns that are populated from your actual column names — no need to remember or retype them — and the converter also guesses sensible defaults by checking common names like ts/timestamp/@timestamp, level/severity/lvl, category/logger/service, and message/msg/event. If a column genuinely shouldn’t map to any of the four core fields, choosing (none) in its dropdown excludes it — the converter won’t silently substitute an auto-detected column back in. Any column not mapped to timestamp, level, category or message is appended to the line automatically as key=value, so nothing in the original row is lost even if you only mapped a couple of fields.

Output matches the style you need: classic bracket lines like [2026-08-15 09:12:33] [ERROR] [api] message, plain space-separated lines, flat key=value lines that keep every column visible, or a fully custom template built from {ts}, {level}, {category}, {message} and {fields} placeholders for matching whatever format a downstream tool expects. Levels are normalized against common aliases (warning→warn, err→error, critical/crit→fatal, and so on) and can be uppercased for output; unrecognized values pass through unchanged rather than being dropped. Timestamps can stay exactly as they were in the CSV, or convert to ISO 8601 UTC or a compact YYYY-MM-DD HH:MM:SS form — values that don’t parse as valid dates are left untouched instead of turning into “Invalid Date”.

• Robust CSV parsing: comma, semicolon or tab separators with correct quoted-field handling, including embedded commas, quotes and newlines.
• Ragged-row tolerant: missing trailing cells become empty strings instead of breaking the conversion.
• Header or headerless: columns identified by name, or by position with a label editor for headerless exports.
• Auto column mapping: timestamp, level, category and message detected from common column-name aliases.
• Manual override: selectors are populated from your real column names, and choosing “(none)” actually excludes a field rather than reverting to auto-detection.
• Bracket output: [ts] [LEVEL] [category] message lines like classic log frameworks.
• Plain output: simple ts LEVEL category message text lines.
• Key=value output: flat lines that keep every column visible and greppable.
• Custom template: compose lines with {ts} {level} {category} {message} {fields} placeholders.
• Level normalization: common aliases collapsed to canonical names, with optional uppercasing.
• Timestamp reformatting: keep original, or convert to ISO 8601 UTC or a compact date-time string, with invalid dates left as-is.
• No data loss: every unmapped column is preserved as a key=value pair on the line.
• Local processing: all conversion happens in your browser, nothing uploaded.

1. Paste your CSV. Drop comma, semicolon or tab-separated data into the input, or click Load sample to see a mixed, slightly ragged export.
2. Set the separator. Match the file you pasted — comma, semicolon or tab — and toggle the header row if your data starts straight into values.
3. Map columns. The converter auto-detects ts, level, category and message columns from common names; adjust any dropdown, or pick “(none)” to leave a field out entirely. For headerless files, fill in the column labels box instead.
4. Pick an output style. Bracket, Plain, Key=value or Custom template, plus timestamp format and uppercase-level options.
5. Convert. Every row becomes a log line, with remaining (unmapped) columns appended as key=value pairs, and a row/byte count shown below.
6. Copy, download or print. Copy the result, download it as a .log file, or print it for a record.

Example 1 — Replaying an audit export. A compliance export is a CSV table of events. Map ts, level and message columns, pick Bracket output, and every audit event is restored to the log format the SIEM expects.

Example 2 — Feeding a test harness. A test fixture stores expected events as CSV. Convert the rows to plain log lines and feed them to the parser under test, comparing output line by line.

Example 3 — Headerless data. A legacy exporter omits headers. Turn the header switch off, set column labels in the label editor, and the rows render with meaningful key=value names instead of col1, col2.

Example 4 — Custom format for a tool. A monitoring script wants lines shaped like user_id={user_id} level={level} message={message}. Use the custom template and render each row exactly as the script parses them.

Example 5 — Merging separate columns. Your export splits date and time into two columns. Map the combined or time column as the timestamp, and every line gets a single, sortable timestamp instead of two separate fields.

Example 6 — Deliberately dropping a column. Your CSV has a category column you don’t want in the output at all. Select “(none)” in the Category col dropdown and it disappears from every line instead of reappearing on the next edit.

• Reverse the pipeline: tables turn back into the text logs systems and humans understand.
• Messy CSV safe: quoted cells, embedded separators, ragged rows and headerless layouts all handled.
• Column mapping made easy: dropdowns built from your real column names, with predictable “(none)” behavior.
• No data loss: every unmapped column stays on the line as key=value.
• Output flexibility: bracket, plain, key=value or a fully custom template.
• Consistent levels and timestamps: normalized aliases and safe date reformatting, with graceful fallback on unparseable values.
• Private by design: data is processed locally and never uploaded.