CSV Import Traps: Leading Zeros, Quotes & Multiline Fields

Inspect a CSV without changing its identifiers

1. Download the synthetic fixture below, then open CSV Viewer and choose that local file. Alternatively, use “Open example in tool” to open the shared example immediately.
2. Use the comma delimiter and keep the header enabled. The three columns are id, title and notes.
3. Inspect the first row. Its identifier is the text 001, its title is Quoted, value, and its notes contain an embedded newline.
4. Inspect the second row. Its identifier is 002 and its title is Escaped "quote".
5. If exporting JSON or JSONL, keep optional type inference off to preserve these identifiers as strings. Preview and export locally, then reopen the result to verify the values.

Expected results

The fixture has three columns and two data rows, even though the first notes field spans two physical lines. Splitting on every newline would break that logical row; splitting on every comma would break its quoted title.

Data Demon Systems maintains this synthetic input alongside the browser's built-in multiline CSV example. Its record expectations follow RFC 4180. An edit uses an overlay and does not overwrite the source file.

Opening a larger CSV

Choose the file rather than pasting the entire document. The viewer reads it locally in chunks and uses virtual rows; search and inspect the records you need before choosing an export path. Keep delimiter and header settings consistent with the source.

The supported source limit is 4 GiB, not a promise that every device processes that size equally quickly. Supporting browsers can stream an export to a file handle; the fallback Blob download is capped at 100 MB. This two-row fixture tests record correctness, not throughput.

The versioned CSV evidence and benchmark method document reproducible inputs. Preserve the browser version, machine details and complete benchmark output with any timing claim.

Limitations

This fixture is UTF-8 CSV with a comma delimiter and a header. Other delimiters can be overridden and reindexed. CSV is not a spreadsheet: formulas, row insertion, global sorting and arbitrary legacy encodings are outside this workflow. Header names must be unique when exporting objects as JSON.

For newline-delimited JSON output, continue with JSONL & NDJSON Viewer Online to inspect the resulting records.

Complete synthetic fixture

id,title,notes
001,"Quoted, value","Line one
Line two"
002,"Escaped ""quote""","CRLF rows remain logical records"