Introduction

Pandora is a universal database proxy and SQL-Database engine, that implements high-performance and security requirements for enterprise analytical applications.

The primary goal of Pandora is to provide a unified (universal) data interface for machine learning- and data analysis applications, to facilitate their integration into existing operational data landscapes. To achieve this goal, Pandora implements the two fundamental layers of a data warehouse:

The integration layer of Pandora utilizes SQLAlchemy to allow it’s connection to a variety of SQL-Databases (e.g. IBM DB2, Oracle, SAP, MS-SQL, MS-Access, Firebird, Sybase, MySQL, Postgresql, SQLite, etc.). Thereupon it provides native support for flat file databases (e.g. CSV-Tables, R-Table exports), laboratory measurements and data generators.

The staging layer of Pandora is implemented as a native SQL-Database engine, featuring a DB-API 2.0 interface with full SQL:2016 support, a vertical data storage manager and real-time encryption. This allows the data analysis application to integrate a variety of different data sources, by using a unified data interface (and SQL dialect).

Pandora is open source, based on the Python programming language and actively developed as part of the Smart Analytics project at Frootlab.