CSV Converter

CSV Summary

{{ metrics.rowCount.toLocaleString() }} rows {{ metrics.columnCount.toLocaleString() }} columns {{ metrics.header ? 'Header detected' : 'No header' }} {{ metrics.delimiterLabel }} Preview trimmed

CSV:

Delimiter:

Custom delimiter:

Quote character:

Escape style:

Header row:

Preview rows:

rows

Trim fields:

Skip empty rows:

Detect data types:

Detect dates:

Empty as null:

XML root tag:

XML row tag:

Row index attribute:

Column naming

Column	Display label	JSON key
{{ base \|\| `Column ${idx + 1}` }} Active key: {{ headers[idx] \|\| baseHeaderKeys[idx] }}
No columns detected yet.

Leave fields blank to reuse detected labels and keys. JSON keys are sanitized and deduplicated automatically.

#	{{ header }}
{{ row.index }}	{{ cell }}
No rows to display.

Showing first {{ max_preview_rows }} rows. Export actions include the full dataset.

Column	Key	Type	Non-empty	Null	Empty	Unique	Sample value
{{ profile.index }}. {{ profile.label }}	{{ profile.key }}	{{ profile.type \|\| 'empty' }}	{{ profile.nonEmpty.toLocaleString() }}	{{ profile.nullCount.toLocaleString() }}	{{ profile.emptyCount.toLocaleString() }}	{{ profile.unique.toLocaleString() }}	{{ profile.sample \|\| '—' }}
No columns profiled yet.

Null counts include explicit null markers. Empty counts represent blank strings after trimming.

{{ jsonArrays }}

{{ jsonLines }}

JSON Lines requires a header row to emit key/value pairs.

{{ xmlText }}

{{ htmlTable }}

{{ markdownTable }}

{{ tsvText }}

{{ jsonRecords }}

JSON objects require a header row. Switch to data mode or set header override.

Tags: Data Format , Devtools , Editor , File Tools , Text

Export to PDF Fullscreen

Embed:

Stylized table grid representing rows and columns.

Comma separated values files are plain text tables that store rows and columns for simple data exchange, and quick conversion and inspection help you move work forward with fewer mistakes.

Paste data or drop a file, choose detection or set the delimiter, then review the live preview and the column profile to confirm structure and types, for example when you convert CSV to JSON Lines for a pipeline.

The tool reads your text, identifies the likely separator, checks for a header, and infers basic types so booleans, numbers, and dates are handled predictably, then it prepares clean outputs you can reuse.

A short sample such as name city score with a few rows will show a detected header, numeric scores, and ready to copy JSON, XML, HTML, Markdown, or a spreadsheet table in seconds.

Valid structure still needs human judgment, so watch the warnings and the profile counts, and set stable labels and keys for consistent exports across repeated runs.

Technical Details:

Comma separated values (CSV) are row oriented text tables built from fields separated by a delimiter and optionally wrapped in quotes. The parser supports comma, tab, semicolon, pipe, or space separators, with either double or single quotes and repeat quote or backslash escapes. Header presence can be detected automatically or fixed by choice. JavaScript Object Notation (JSON) is emitted as records or arrays, JSON Lines as one object per line, Extensible Markup Language (XML) with customizable root and row tags, and Tab Separated Values (TSV) and HTML or Markdown tables are also generated.

Delimiter detection scores several candidates using the average number of fields per non empty line and a small penalty for inconsistency between lines. The highest score wins and then parsing proceeds with the chosen delimiter and quoting rules.

Column profiling reports the dominant kind in each column and counts non empty cells, nulls, empties, unique values, and a sample. Type inference recognizes booleans, numbers including scientific notation, the words null and undefined as null, and optionally ISO like dates that are normalized to an invariant form.

Results are comparable when headers and keys stay stable across runs, so set overrides when a header is missing or duplicated. Keep in mind this is a text parser tuned for pragmatic CSV conventions, and ambiguous inputs may need manual settings for best outcomes.

\begin{matrix} s (d) & = & \overline{n} - 0.1 \cdot V \\ \overline{n} & = & \frac{\sum n_{i}}{m} \\ V & = & \sum | n_{i} - n_{i−1} | \end{matrix}

Symbols and units
Symbol	Meaning	Unit/Datatype	Source
d	Delimiter candidate	string	Derived
nᵢ	Fields counted in line i	integer	Derived
n̄	Average fields across sample	real	Derived
V	Sum of absolute line to line differences	real	Derived
s(d)	Delimiter score	real	Derived
q	Quote character	" or ' or none	Input
e	Escape style	double or backslash	Input
H	Header detected	boolean	Derived

Worked example

name,city,score
Ada,London,98
Grace,New York,91

\begin{matrix} s (,) & = & 3 - 0.1 \cdot 0 = 3.0 \\ s (⇥) & = & 1 - 0.1 \cdot 0 = 1.0 \end{matrix}

Comma wins, the first row is treated as a header, scores are parsed as numbers.

{"name":"Ada","city":"London","score":98}

<row index="1"><name>Ada</name><city>London</city><score>98</score></row>

Validation and bounds
Field	Type	Min	Max	Step/Pattern	Error Text
Upload	.csv .tsv .txt; text/csv; text/tab-separated-values	—	—	accept attribute	—
Preview rows	number	10	1000	step 10	—
Custom delimiter	text	0	4	max length 4	—
XML root tag	text	0	32	sanitized, leading digit prefixed	—
XML row tag	text	0	32	sanitized, leading digit prefixed	—
Parser	stateful scan	—	—	quotes and escapes honored	Reached end of file while inside quotes.
Detector	heuristic	—	—	5 candidates scored	Unable to resolve delimiter.
Sanity checks	row width	—	—	pad missing cells	Row N has M fields; Irregular row lengths detected.
Header keys	sanitized	—	—	deduplicated	Header duplicated or missing, keys adjusted.

Processing pipeline

Normalize line endings to a single newline.
Score candidate delimiters using field counts across up to 20 lines.
Parse rows with quoting and escape rules.
Decide header presence by heuristics or explicit choice.
Build safe display labels and unique JSON keys.
Coerce values to boolean, number, null, or date when enabled.
Compose preview, profiles, and export formats.
Generate downloads or copy to clipboard on demand.

I/O formats and encoding

Input and output formats
Input	Accepted Families	Output	Encoding/Precision	Rounding
CSV text or file	.csv, .tsv, .txt; text/csv; text/tab-separated-values	CSV, TSV, XLSX, DOCX	UTF‑8 text, binary blobs for spreadsheets	none
—	—	JSON arrays and records, JSON Lines	Pretty printed JSON; one object per line for JSONL	n/a
—	—	XML	Escapes & < > " '	n/a
—	—	HTML, Markdown	HTML entity escaping; Markdown pipes escaped	n/a

Networking and storage

Processing is browser based. A spreadsheet library is fetched from a public content network on first XLSX export; all parsing, profiling, and text conversions run locally thereafter.

Performance and complexity

Delimiter scoring examines up to 20 lines per candidate, and parsing is linear in the size of the input. Profiling uses bounded sets for uniqueness and simple counters per column.

Diagnostics and determinism

With identical inputs and options the outputs are identical. Warnings are deduplicated, and summary badges reflect the three most common column kinds.

Security considerations

HTML and XML outputs escape special characters, JSON uses a safe serializer, and JSON keys and XML tags are sanitized and deduplicated. Clipboard writes and downloads occur only after explicit user actions.

Assumptions and limitations

Space as a delimiter can be ambiguous for text with multiple spaces Heads‑up
Header detection is heuristic and may need manual override
Date detection only considers ISO like forms when enabled
Numbers use a dot decimal separator and no thousands separator
Very wide or long files are limited by device memory
XML tag names are sanitized and may be altered
Record oriented JSON requires a header to supply keys
Spreadsheet export depends on a third party script load

Edge cases and error sources

Quoted fields with unmatched quotes
Backslash escapes inside quoted text
Scientific notation and leading plus signs in numbers
Empty strings versus explicit null markers
Irregular row widths requiring padding
Duplicate or blank headers producing adjusted keys
Keys beginning with digits prefixed to stay valid
Non ASCII characters that need escaping in HTML or XML
Large unique value sets truncated in profiling
Embedded delimiter characters inside quotes

Privacy & compliance

Files are processed locally; nothing is uploaded. One optional script is fetched to enable spreadsheet export.

Step‑by‑Step Guide

CSV conversion and profiling follow a short sequence to check structure and export clean formats.

Paste CSV text or drop a file Input.
Pick a delimiter or use automatic detection Delimiter.
Confirm header treatment and set label or key overrides if needed Header.
Enable type and date detection when appropriate Typing.
Review the profile and warnings, then choose an output tab and copy or download.

Example: a three column grade sheet with five rows becomes JSON records with keys name city score and a spreadsheet you can open immediately.

When the preview matches expectations, your exports will mirror it.

FAQ

Is my data stored?

No. Parsing, profiling, and conversions run locally. A single external script may be fetched to enable spreadsheets. Clipboard and downloads require clicks.

Avoid pasting secrets into shared devices.

Why are JSON records disabled?

Record shaped JSON and JSON Lines need column names. Turn on header mode or add label overrides so keys exist before exporting.

How accurate is delimiter detection?

It favors a candidate that yields many fields with low line to line variation. Ambiguous data may need a manual choice for reliable results.

What formats can I export?

CSV, TSV, spreadsheet and document tables, JSON arrays, JSON records, JSON Lines, XML, HTML, and Markdown tables are available.

Can I use it without a network?

Yes for text formats. Spreadsheet export needs one library fetched once. After that, conversions continue locally.

How do I convert CSV to XML?

Paste data, confirm header and keys, open the XML tab, adjust root and row tags, then copy or download the result.

What does “Preview trimmed” mean?

Only the first N rows show in the preview for speed. Exports always include all rows unless you remove them from the input.

Is there a cost or license?

Use is unrestricted here. Check the parent site for any broader licensing notes.

Glossary

CSV: Comma separated values text table.
Delimiter: Character that separates fields in a row.
Quote character: Character that encloses fields containing delimiters.
Escape style: How quotes and special characters are represented.
Header row: First row containing column names.
JSON: JavaScript Object Notation data format.
JSON Lines: One JSON object per line text stream.
TSV: Tab separated values text table.
XML: Extensible Markup Language markup format.
Null: Explicit missing value marker.