Mixed-Source Computing：SPL Practices

The problem of mixed data source computations arises from the diversity of data sources, including relational databases, NoSQL databases, cloud storage, APIs, and file systems. Logical data warehouses can facilitate mixed-source computations to some extent, but they are often heavy and complex, making them suitable only for large-scale scenarios. SPL provides a lightweight and real-time mixed-data source computing capability, enabling mixed-source computations on any accessible data sources. SPL has two types of data source connectors: native connectors and external connectors, which support a vast variety of data sources. SPL supports and encourages using a data source's native syntax to access and compute data, and offers supplements if the data source's computing ability is insufficient. SPL provides two types of data objects for accessing data from sources: table sequence and cursor, which correspond to in-memory data table and streaming data table, respectively. Unlike logical data warehouses, SPL does not require pre-defined metadata for mapping, and accesses data directly using methods provided by the data source. SPL can be used in IDE for configuring data sources, and in applications for integrating with the application and invoking SPL scripts using JDBC. SPL is open-source and supports a wide range of data sources, making it easy to implement mixed computations.