• Overview
• Parquet Files driver connection settings
• Features and capabilities
  • Advanced SQL query capabilities
  • Folder structure
  • Internal database
  • Additional features

This guide provides instructions on how to set up and use Parquet files with DBeaver. The Parquet driver allows you to work with Parquet data as if it were in a database. You can retrieve data and apply filters, sorting, and other operations, even combining data from multiple files.

Before you start, you need to create a connection in DBeaver and select the appropriate Parquet driver. If you haven’t done this, see our Database Connection article.

Important: When using the Parquet driver, all connected Parquet files are read-only. To make changes, you need to update the original files outside DBeaver.

This section describes how to set up a connection using the Parquet driver. The connection settings page requires the following fields:

Tip: When using the Folder option, DBeaver scans the directory up to two levels deep for Parquet files. For more information, see folder structure. If you select a folder, DBeaver organizes files in schemas based on their directory structure.

The Parquet driver supports the full range of SQL queries:

• Simple queries (e.g., SELECT * FROM table): Data is read directly from the Parquet file.
• Complex queries (e.g., using WHERE, JOIN, ORDER BY, GROUP BY): When a complex query is executed for the first time, the driver imports the entire Parquet file into an internal database to enable advanced SQL functions. Subsequent queries run faster because the data is already imported into internal database.

When working with a folder containing multiple Parquet files, DBeaver organizes them as follows:

If your folder looks like this:

Data/
├── employees.parquet
├── sales.parquet
└── Reports/
    └── monthly.parquet
    └── yearly.parquet

DBeaver will create:

• Default schema: employees, sales
• Reports schema: monthly, yearly

Tip: To focus on specific files, consider selecting individual files or folders when configuring the connection.

When you execute a complex query (such as WHERE, JOIN, GROUP BY, or ORDER BY.), on a Parquet file for the first time, the Parquet driver processes the data by importing it into a temporary internal SQLite database.

By default, this internal database stores data temporarily on the disk during your session and is cleared when DBeaver restarts. To speed up queries on the same file in future sessions, you can specify the internalDbFilePath option in the Driver properties tab (e.g., C:\User\database.db) to reuse the processed data.

DBeaver provides additional features compatible with Parquet driver, but not exclusive to it: