Can I query a Parquet file without Python or Spark?

Yes. Drop a Parquet file into SQL on Your Files and run SQL on it in your browser with DuckDB.

Does it handle large files?

Up to roughly 100 MB of CSV works normally, and larger files can use large file mode to convert, query, profile or split them. It depends on your computer memory.

Free Data Tools for Data Engineers

Convert JSON, CSV, Excel and Parquet, run SQL on files with DuckDB, profile data, compare snapshots and split large files, all in your browser.

The problem

Data engineers spend a lot of time on small file jobs around the pipeline: convert a vendor file, look inside a Parquet file, check a column, compare yesterday with today. Spinning up code or a cluster for each one is slow.

Typical jobs and the tool for each

Convert: JSON, CSV, Excel and Parquet in any direction, one file or a whole folder into a ZIP, with nested JSON flattened into columns.

Query: SQL on Your Files runs SELECT, JOIN and GROUP BY on CSV, Excel, JSON and Parquet in the browser with DuckDB.

Check: Data Profiler reports types, missing values, duplicates and statistics, and Data Comparison compares two snapshots by a key, which gives a CDC-style list of inserts, updates and deletes.

Shape: Data Transformation joins, filters and cleans, SQL Generator writes CREATE TABLE and INSERT scripts for five databases, and Merge or Split Files handles Parquet and CSV.

Trace: Data Lineage records which tool changed which file.

Limits to know

The tools use your browser memory. As a guide, CSV files up to about 100 MB work well, JSON up to about 50 MB, Parquet about 30 MB and Excel about 20 MB. Larger files can use large file mode, which can convert, query with SQL, profile or split them. This is not a pipeline scheduler, and there is no API or command line.