How it works
Add your JSON Lines / NDJSON file. MimiFile checks the format, converts it to Apache Parquet, and provides a temporary download when the result is ready.
Convert JSONL / NDJSON records to Parquet when logs, events or exported records need a typed, compressed columnar format for analytical queries or storage.
JSONL → PARQUET
Drag and drop, paste a file, or browse your device.
Files are temporary and deleted automatically.
Limits: 50 MB per file · 128 MB total per batch.
Add your JSON Lines / NDJSON file. MimiFile checks the format, converts it to Apache Parquet, and provides a temporary download when the result is ready.
MimiFile reads JSONL into Arrow record batches and writes Parquet with Zstandard compression, dictionary encoding and statistics. Validate the inferred schema before production use.
Current source limit : 50 MB per file. Up to 10 files can be submitted in a batch, subject to the global batch ceiling.
JSONL is practical for append-only logs, event streams and record exports. Parquet is optimized for typed columnar storage and analytical reads, making it a better fit for many query and data-lake workflows.
The data worker writes Parquet 2.6 using Zstandard compression, dictionary encoding and column statistics. The result is a real columnar Parquet file, not JSON wrapped in a different extension.
JSONL schema is inferred from the data read by the worker. Check column types, nullability, timestamps, nested fields and schema consistency before relying on the output in a production pipeline.
When a tabular destination cannot represent lists or objects directly, nested values are serialized as JSON text where supported instead of being silently discarded.
Nulls and timestamps are preserved when the destination supports them. Binary values can be represented as Base64 text in text-based or spreadsheet outputs.
Check column types, null values, timestamps and nested fields before using the converted file in a production data pipeline.