Turn JSONL data files into PARQUET. It is converted in your browser, so the file never leaves your device.
Drop JSONL files here
or choose from your computer paste with Ctrl V
from Photos or Files
Ready · other file types work too0 bytes uploaded, ever
Drop JSONL (NDJSON) files onto the box above, or click it to choose them. ToAny reads the file's contents to detect the type, so a wrong extension does not matter.
Press Convert. There is nothing else to set. The work happens in this tab.
Save the PARQUET, or download everything as one ZIP if you converted several files.
The JSONL to PARQUET converter is part of this page and runs inside your browser. Your data files are never sent to ToAny or to anyone else.
Nothing has to travel to a server and back, so the conversion starts the moment you press Convert, even on a slow connection.
No sign-up, no email address and no watermark. Add several files and download them together as a ZIP.
JSONL is not a single valid JSON document, so ordinary JSON tools reject it. Parquet is compact, typed and fast to analyse. JSONL is at home in training data, logs and record streams processed line by line; PARQUET is the usual choice for large analytical datasets.
Each line must be one complete JSON value and blank lines are skipped; if a line is not valid JSON the conversion stops and the message names its line number. Each column gets one type chosen from all of its values: a column is numeric only if every non-empty cell is a number, and anything mixed, or a value such as a ZIP code with a leading zero, stays text.
JSON Lines holds one JSON value per line, so a file can be read or appended to a record at a time. It is the usual format for machine-learning datasets and logs. It is the same convention as NDJSON; only the name and the file extension differ.
Opens in Text editors, jq, pandas and most data tools
JSONL converterParquet stores a table column by column, with the type of every column recorded in the file and each column compressed on its own. That makes it small and quick to scan for analysis, which is why data lakes and warehouses use it. It cannot be read in a text editor or a spreadsheet.
Opens in pandas, Spark, DuckDB, BigQuery, Snowflake and other analytics tools; not Excel
PARQUET converter| JSONLJSON Lines | PARQUETApache Parquet table | |
|---|---|---|
| Stands for | JSON Lines | Apache Parquet |
| Type | Plain-text records | Columnar data file |
| Compression | None | Lossless (per column: Snappy, Zstandard, gzip and others) |
| Introduced | The early 2010s, by no single standards body (jsonlines.org) | 2013, by Twitter and Cloudera, now the Apache Software Foundation |
| Opens in | Text editors, jq, pandas and most data tools | pandas, Spark, DuckDB, BigQuery, Snowflake and other analytics tools; not Excel |
| Best for | Training data, logs and record streams processed line by line | Large analytical datasets |
| File extension | .jsonl | .parquet |
No. The JSONL to PARQUET conversion runs inside your browser, on your own phone or computer. The file is not sent to ToAny or to anyone else.
The values are kept. Anything PARQUET has no way to express, such as formatting, formulas or nesting, is simplified or left out.
Yes. Add as many as you like, convert them together and download each result or all of them as one ZIP.
The steps are the same everywhere because ToAny runs in the web browser: open this page in a current version of Chrome, Edge, Safari or Firefox, add the JSONL file, press Convert and save the PARQUET. There is nothing to install.
Yes. Data files up to 300 MB are accepted; a larger file is refused with a message. Beyond that the limit is your device's memory, since nothing is uploaded; a computer copes with large files better than a phone.