Move spreadsheet exports into JSON pipelines without a desktop ETL step
Analysts live in Excel and Google Sheets while platforms ingest JSON and Parquet. CSV sits in the middle as the lowest-common-denominator handoff — and also the format most likely to break on commas inside quotes, locale-specific decimals and ragged rows. This converter reads pasted or uploaded CSV in the browser and emits JSON arrays of objects, or reverses the direction when you need a flat file from an API sample. Processing is local so partner extracts with customer identifiers never pass through a server-side upload form.
CSV ⇄ JSON workflow
- Confirm delimiter, quote character and whether row one is headers.
- Paste or load CSV, then inspect the JSON preview for type surprises.
- Format JSON for readable review and validate syntax.
- Generate SQL or load to SQLite when the shape looks stable.
Delimiter and encoding pitfalls
European exports sometimes use semicolon separators while US files default to comma. UTF-8 with BOM still appears in finance downloads — if characters look wrong, re-export from the source tool with explicit encoding. Empty trailing columns and all-null rows show up often in manual spreadsheets; drop them before you treat distinct counts as gospel. When text fields contain newlines, ensure quoting survived the export or rows will shift.
Downstream JSON tooling
After conversion, prettify with the JSON Formatter, validate with the JSON Validator, compare supplier revisions in JSON Diff. Build INSERT or CREATE statements with the CSV → SQL Import Helper when you are headed for a database proof of concept. For odd delimited text that is not quite CSV, normalize through the Text ⇄ CSV Converter before you land here.
Handoff hygiene
Agree on date formats and null sentinels with suppliers before automating loads. Keep a hashed sample row count in your ticket so reviewers know whether the paste is complete. Redact PII before sharing JSON even locally in screen shares. When configs are YAML, use the YAML ⇄ JSON Converter for service settings and reserve this page for tabular extracts.
Headers, types and export hygiene
The first row is treated as column names unless you override that behaviour in the tool. Numbers may arrive as strings when leading zeros matter — cast explicitly in your target system. After conversion, spot-check null sentinels such as empty strings versus literal "null" text. Large spreadsheets should be sampled before you paste entire exports into the browser.