When you need Parquet rows in something a script or an API can consume, JSON is the obvious target. This converter decompresses the file, converts every Parquet type into its JSON equivalent, and can rebuild nested objects from flattened column names.
Reading Parquet
Auto-detect reads the first rows and picks the separator that splits them consistently.
Turn off when the file starts straight into data. Columns are then named column_1, column_2, and so on.
Turns "42" into a number and "true" into a boolean. Values with leading zeros stay text so IDs and zip codes survive.
Which worksheet to convert. Filled in once the workbook is read.
Dotted path to the array of records, such as data.items. Leave empty to use the whole document.
Turns {"user":{"id":7}} into a user.id column. Arrays are kept as JSON text.
Pick the encoding the file was written in. Wrong encoding shows up as garbled accents.
Comma-separated words that should become null instead of text.
Writing JSON
Snappy is the default across Spark, DuckDB and pandas. Gzip is smaller but slower to read.
Turns a user.id column back into {"user":{"id":7}}.
Excel needs a byte order mark to open UTF-8 files with accents correctly.
Rows per row group. Larger groups compress better; smaller groups let readers skip more.
conversion-preroll
This short ad is what keeps the converter free.
Inspecting Parquet records by eye?Query Parquet without exporting with ParquetReader — opens Parquet files directly.Typed, compressed, columnar storage built for analytical scans. Nested, self-describing records that every language can read.
| Parquet | JSON | |
|---|---|---|
| Type information | Full — ints, floats, booleans, timestamps, decimals | Strings, numbers, booleans, null |
| Structure | Columnar, with nested type support | Nested objects and arrays |
| Compression | Built in (Snappy, Gzip, Zstd) | None |
| Typical file size | Typically 5–15% of the CSV | Largest — keys repeat on every record |
| Schema | Stored in the file footer | Implicit, per record |
| Human-readable | No — binary | Yes, though verbose |
| Tool support | Spark, DuckDB, pandas, BigQuery, Athena | Every language, natively |
Parquet distinguishes INT32, INT64, FLOAT and DOUBLE. JSON has only "number", backed by a 64-bit float. Integers above 2^53 are emitted as strings here rather than rounded, which is the safe choice but does change the type on the way out.
Every key is repeated on every record and all compression is gone. A 50 MB Parquet file can easily become 700 MB of JSON, so consider JSON Lines if the result is going into a pipeline rather than an editor.
JSON has no date type. Timestamps are written as ISO-8601 text, which every language can parse but which no longer carries the timezone precision the Parquet column may have had.
Most application code reads JSON natively and has no Parquet support at all. Converting once is simpler than adding a Parquet dependency.
Reading a handful of rows as formatted JSON is the fastest way to see whether a transformation did what you expected.
MongoDB, Elasticsearch and similar systems import JSON directly.
sticky-bottomYour file is downloading
post-conversion